Galapagos tortoise.

A behavioural scientist at an American zoo had a problem. Each day she had to arrange for the transfer of the giant Galapagos tortoises from their overnight heated enclosure to an open air garden where they were on public display.

Carrot or stick?
Animal behaviourist Heather Simpson explores whether your dog will respond best to reward or punishment.

But because each tortoise weighed an average 91kg (200Ib), several members of staff had to lift the tortoises on to trolleys so that they could be wheeled to and fro each morning and evening - a time-consuming business.

But then the scientist thought: "What if I could train the tortoises to voluntarily walk up and down the path - how convenient that would be." And so she did! Using positive reinforcement, she
taught the tortoises to touch a 'yellow lollipop' in return for a food reward of cactus stems and by training them to walk forward to touch the 'lollipop', the tortoises learned to walk from one enclosure to the other. After only a few ten-minute sessions, the tortoises were walking more than nine meters (30ft) - and, remember, we are talking cold blooded reptiles here.

Nice or nasty?
Using positive reinforcement to train animals is certainly not new. In the 1900s a psychologist called
Edward Thorndike documented positive reinforcement for the first time and in the 1930s, behavioural psychologist B.F. Skinner (and others) expanded that work by proving that there are just four types of reinforcement that can cause an animal (or person) to change his behaviour:

All four reinforcers work.
One of the first important things to understand is that, when appropriately applied, all the four reinforcers work. Skinner proved the point many times and he did so with many species of animals from raccoons to dogs and even chickens. So, in many respects, whether I choose to train my dog using positive reinforcement or punishment is actually down to how I might feel morally about the choice. Not entirely, of course; if the welfare of the dog was to be severely compromised through my choice of punishment, then the dog would be protected by law.

But putting such extremes to one side, all dog trainers have to consider ethics. It may be effective to train my dog to perform perfect recalls by not feeding him before I take him out for a walk, that is, to use negative punishment, but is it ethical? Behaviourists recognize that to rely on the assumption that any type of training is acceptable as long as it is legal, is not automatically good for the dog. This is because dogs, just like us, experience different emotions and levels of stress depending on which reinforcer is used.

Even Skinner, when he realized the power of all four reinforcers, began to consider the consequences of their use on the animal. What does it feel like to be at the receiving end of punishment or negative reinforcement? Is there an adverse consequence for the dog if we use these three unpleasant reinforcers often?

I'll explore these consequences in more detail over the page but for now it is worth remembering that many scientists now say that the main issue is not that all four reinforcers work. What is important is that, as animal trainers, we recognize we have an ethical responsibility to stay in the positive reinforcement box when we train our animals. If we do not, we adversely affect our relationship with the dog and, in some cases, can even create health problems caused by chronic stress.

As trainers, we are responsible for ensuring our dog understands what we are asking of him. If you take the time to become excellent at setting up requests correctly, you will find that your dog will learn to readily understand your cues and, because he starts to offer you the behaviour you always want, it becomes easy to use positive reinforcement - always.

Be a bringer of good things If you want a relationship with your dog where he really wants to be with you, adding nasty things to his environment or withholding things he really wants may get you the behaviour you want in the short term but it does little to create the trust that most people want from their dog.

We all know that dolphin and seal trainers can teach these animals to perform complicated behaviours for a fish reward. Many enlightened zoos have switched completely to the use of positive reinforcement – famous examples include training a pregnant rhinoceros to stand quietly to be scanned, or teaching giraffes to raise their feet to have their hooves trimmed. Once you become aware of the awesome possibilities of positive reinforcement training, you will start finding examples everywhere. If a zoo animal trainer can teach tigers to swim with him for fun, then surely we dog trainers can teach a dog to come back to us without resorting to punishment by keeping him hungry first? Why only positive reinforcement? Consider this situation - a woman would like her partner to mow the lawn. He would rather watch the football on TV. Because the woman wants to increase the chances of the behaviour happening, she has only two options available, positive or negative reinforcement. So she tries the following: "Honey, if you mow the lawn this afternoon, I will bake your favourite cake for you," (positive reinforcement). Or the negative approach:
"Mow the lawn!" Then a few minutes later: "Mow the lawn!" And then again: "Mow the lawn! I am going to nag and nag until you do," (negative reinforcement).

I n both cases, the man gets off the couch and mows the lawn. His behaviour (he mows the lawn) looks identical but see how differently he would be feeling inside. Enthusiasm is the big difference between positive and negative reinforcement. Using negative reinforcement, you can undoubtedly get your dog to do as you want but the feeling inside him is very different.

If he starts to learnthat you will only use positive reinforcement, he starts to become more trusting and more enthusiastic as he tries harder to give you the behaviour that gives him a reward. People who only use positive reinforcement with their dogs (no ignoring, no withholding of food, no gadgets designed to inhibit or constrain) find that if they do not close a training session appropriately, their dog will carryon without them - jumping jumps, climbing on boxes, pushing balls and all just for the enthusiasm and chance of a reward.

Don't be a bully.
Stopping the negative reinforcement does not make it positive reinforcement for the dog.
Just because the woman stops nagging when her partner finally goes out and mows the lawn, it is not a positive experience for the man - it's just the absence of the nagging. Positive reinforcement has to involve a positive benefit - both for humans and dogs!

So much of dog training is built around putting pressure on the dog to get him to perform and then releasing that pressure only when he does so; or withholding something he really wants until he does as we ask. People are, in effect, saying to their dogs: "You have to do this - or else". This is a domineering, bullying approach. But many trainers try to use unpleasant reinforcers and then follow it up with a food treat. In many cases, this technique does work but it gets back to how the dog is feeling inside. Let's take our analogy of lawn mowing. If the woman relentlessly nagged the man but then baked him a cake later, it might improve their relationship a little but the fact would remain that the man endured an unpleasant bout of nagging first.

So do not be taken in by trainers who suggest that you use a negative punisher such as ignoring the dog and then follow this up with a treat when you get him to do as you want. Human psychologists have proved that being ignored by a partner who is sulking can be one of the most stressful negative punishers and its repeated use by one partner is an important factor leading to a breakdown in the relationship.

Food rewards.
If we use a food reward, we often find ourselves accused of using bribes but here is the big difference: if you work all month, do you feel bribed when you get your salary? For positive reinforcement to work, we have to give the dog a reward that makes sense to him. If I said to my dog: "I'll give you a hundred quid if you run back to me when I call you," would it make a difference? Of course not. We have to use rewards that are meaningful to dogs. Behavioural scientists call this saliency.

Dogs that are unmotivated in training are invariably showing this behaviour because the saliency of the chosen reinforcer is wrong. Some dogs are motivated to learn more by play than by food; others are motivated by your attention, such as stroking. It is important that you work out which positive reinforcer is meaningful for your dog. It is also important for you to realize that, as the training progresses, the timing and variation of the reward needs to change.

If I keep feeding my dog the same biscuits in every training session, he will soon find this a turn-off. Owners often call this boredom but scientists call it asymptote because there is a crucial difference of understanding. Asymptote is the point the dog reaches where he stops trying because he has had enough of the reward - because it is always the same and completely predictable. If you have experienced asymptote in your dog, you may have noticed that when he is off-lead, he walks away from you or goes and lies down. Change the reinforcer from a dry biscuit to a piece of smoked sausage and asymptote disappears for most dogs and they will bound back enthusiastically, ready to do more. At this stage, don't worry that you will be handing out treats forever
- I will help you to change the use of treats to another positive reinforcer, such as stroking, later in the series.

_{source Your Dog Mag.}