Teaching a pet dog complex tasks has a lot of benefits. Often we choose breeds that come from working backgrounds, but may not get the chance to fully express the desire to work or hunt as a pet. So teaching complex tasks can give the dog satisfaction of their working or hunting instinctive drives. When these instinctive drives are satisfied, the dog can be calmer, more aware and emotionally balanced. Also when instinctive drives are satisfied, the dog can be more amenable to other obedience training and less likely to engage in nuisance behaviours.
Back Chaining is the most efficient technique used to teach complex tasks to a dog. When understood, it can also serve as a diagnostic guide to training errors and specific areas where the dog may have put a pattern together that we hadn't anticipated.
The basic idea is that a complex behaviour is broken down into sections and mapped on a timeline. Then each specific section of the behaviour is trained in reverse sequence starting with the last section and working backwards through the timeline.
So, if you want to teach a dog to find and bring your slippers, for example, the timeline might look like this:
1. You give the cue 'slippers'
2. The dog goes to your bedroom
3. The dog opens the wardrobe door.
4. The dog picks up the slippers.
5. The dog holds the slippers in it's mouth.
6. The dog closes the wardrobe door.
7. The dog returns to where you are with the slippers in it's mouth.
8. The sits in front with the slippers in it's mouth until you take them from him.
9. You reward the dog.
Depending on the size of dog and the type of slippers, the dog might only be able to carry one at a time - so there could be more steps in this process.
Each of the above steps will probably need to be broken down into segments depending on the dog, it's training history and communication bond with you.
In the above example, the training starts with rewarding the dog for holding the slippers in its mouth while sitting in front of you and ends with you teaching the cue 'slippers'. The reason why the cue taught from the beginning is that the dog may go through times of confusion or frustration while he's learning and these, if not managed, can lead to him losing interest or engagement in the process. So if the cue for the behaviour is saved for last, when the dog is enthusiastically and efficiently putting all the steps together, the purity of the classical conditioning (reliable anticipation of reward/enthusiasm) of the cue is protected. Classical conditioning is about how the dog feels when he hears the cue.
Teaching the last behaviour first protects the dogs enthusiasm and drive to move toward the behaviour that will deliver the reward. So after the dog is reliably predicting reward for holding slippers in his mouth and allowing you to take them from him, it becomes a lot easier to move to the next step of perhaps placing the slippers on the floor near you and encouraging him to pick them up himself and hold them until you take them and in this way, protect his enthusiasm as you back chain through each step.
An example of how a dog might naturally back chain:
Lets say your dog really likes going for walks. His initial reliable predictor that you will be taking him for a walk is that you are opening the front gate when he is wearing his collar and lead attached. After a period of time and a number of repetitions, you attaching the lead will become a reliable predictor that you are taking him for a walk. Then later, you picking up his lead will be a reliable predictor that you are taking him for a walk. Then changing your shoes before you pick up his lead will become a reliable predictor that you will be taking him for a walk. So the dogs anticipation and enthusiasm starts to build at earlier and earlier cues in the sequence as he puts the patterns together.
Since dogs naturally back chain, and in the case of unwanted behaviours or responses to behavioural cues, knowing how a dog back chains and the types of things that can become cues to behaviours helps to iron out any training or behavioural issues.
Dogs, and people I think, back chain on both positive and negative experiences.
Of course, the complex task can be a game that has a high degree of complexity and one that creates problems for the dog to solve. For example, my last dog was a hunting breed, but she wasn't allowed to hunt animals, and she wasn't really interested in fetching balls, so I taught her a version of 'hide and seek' where I'd put her on a sit-stay somewhere where I could disappear out of view, then I'd go and hide one of her fluffy toys (how and where I hid the toys grew in complexity as time went on). The key that I'd protected the classical conditioning of the cue was that I could put her on a sit-stay, walk out of sight with her toy, but her eye's didn't light up or body tense in anticipation until I said 'where is it?'. Another reason that protecting the cue is important is that if she were around children that had fluffy toys, she was less likely to tense up in anticipation unless she heard my cue, but if she did she was also aware and amenable to my direction to leave off because her hunting instincts were regularly satisfied.
The next thing is that you want the cue to a behaviour to be something that you can control and is explicitly reserved for the behaviour in question. Here's an example that appeared on my newsfeed this morning:
"A common mistake when teaching/training using food.
**********************************************
How many that are training their pups with treats are using the opening of a plastic bag or container as a cue that a treat is available?
I see this mistake so often. The pups owner decides to do a little training session with their pup. They go to the kitchen and open a plastic bag or container. Or open a container just before they start their training session. The pup comes running, 100% focused..... Training session over, close container or bag, and put it away. Most pups will very quickly switch on to this routine.
When using food to train, only the behaviour/response you are teaching should be marked with a sound or word. Don't mistakenly telegraph to your pup (or dog) that food is now available. If you are using treats, you should carry the treats with you, and randomly mark and reward responses when your pup is least expecting it. Don't go to the kitchen and open a container that cues the dog whats coming next, or open a container just before you give a command. This is one reason many that are training pups with food cannot get their pups attention unless the pup is triggered by a sound that food is now available. Carry food around with you in a treat pouch, so your pup never knows when the treat is available; not until a required behaviour or response is marked with a sound or word. Too many are unintentionally cuing the pup before the required behaviour, with a sound that food is coming. - Mark Singer, Dog Behaviourist"
The problem with the opening of cans, jars or packets cuing a dog that a treat is coming is obvious - they can start harassing people for food, or running into the kitchen when you're trying to cook a meal creating tripping or spilling hazards. So apart from Marks suggesting of keeping a treat pouch on you in the early days of training, you could also teach a training cue. I used the words 'are you ready?' for my dogs training cue. Using this meant that I could engage my dog in training without having treats or food on my person and that I could go to a packet of food to get a treat for her - in other words, the presence of food or the sound of packets or jars was NOT a reliable cue that she would get food. The most reliable thing that predicted reinforcement for her was my words.
Here is a 'dancing with dogs' routine being trained by back chaining. In this example, the clicker is a 'reward marker'. It both signals to the dog that it has done the right thing, and lets the dog know that reinforcement is on it's way.
Here is a complete dancing with dogs routine - probably one of the most famous:
Dogs are just as happy to be given tasks to do as a member of their human pack and I think it's good and healthy for them so long as the task is within their abilities, level of health, natural energy levels and resources.
Back Chaining is the most efficient technique used to teach complex tasks to a dog. When understood, it can also serve as a diagnostic guide to training errors and specific areas where the dog may have put a pattern together that we hadn't anticipated.
The basic idea is that a complex behaviour is broken down into sections and mapped on a timeline. Then each specific section of the behaviour is trained in reverse sequence starting with the last section and working backwards through the timeline.
So, if you want to teach a dog to find and bring your slippers, for example, the timeline might look like this:
1. You give the cue 'slippers'
2. The dog goes to your bedroom
3. The dog opens the wardrobe door.
4. The dog picks up the slippers.
5. The dog holds the slippers in it's mouth.
6. The dog closes the wardrobe door.
7. The dog returns to where you are with the slippers in it's mouth.
8. The sits in front with the slippers in it's mouth until you take them from him.
9. You reward the dog.
Depending on the size of dog and the type of slippers, the dog might only be able to carry one at a time - so there could be more steps in this process.
Each of the above steps will probably need to be broken down into segments depending on the dog, it's training history and communication bond with you.
In the above example, the training starts with rewarding the dog for holding the slippers in its mouth while sitting in front of you and ends with you teaching the cue 'slippers'. The reason why the cue taught from the beginning is that the dog may go through times of confusion or frustration while he's learning and these, if not managed, can lead to him losing interest or engagement in the process. So if the cue for the behaviour is saved for last, when the dog is enthusiastically and efficiently putting all the steps together, the purity of the classical conditioning (reliable anticipation of reward/enthusiasm) of the cue is protected. Classical conditioning is about how the dog feels when he hears the cue.
Teaching the last behaviour first protects the dogs enthusiasm and drive to move toward the behaviour that will deliver the reward. So after the dog is reliably predicting reward for holding slippers in his mouth and allowing you to take them from him, it becomes a lot easier to move to the next step of perhaps placing the slippers on the floor near you and encouraging him to pick them up himself and hold them until you take them and in this way, protect his enthusiasm as you back chain through each step.
An example of how a dog might naturally back chain:
Lets say your dog really likes going for walks. His initial reliable predictor that you will be taking him for a walk is that you are opening the front gate when he is wearing his collar and lead attached. After a period of time and a number of repetitions, you attaching the lead will become a reliable predictor that you are taking him for a walk. Then later, you picking up his lead will be a reliable predictor that you are taking him for a walk. Then changing your shoes before you pick up his lead will become a reliable predictor that you will be taking him for a walk. So the dogs anticipation and enthusiasm starts to build at earlier and earlier cues in the sequence as he puts the patterns together.
Since dogs naturally back chain, and in the case of unwanted behaviours or responses to behavioural cues, knowing how a dog back chains and the types of things that can become cues to behaviours helps to iron out any training or behavioural issues.
Dogs, and people I think, back chain on both positive and negative experiences.
Of course, the complex task can be a game that has a high degree of complexity and one that creates problems for the dog to solve. For example, my last dog was a hunting breed, but she wasn't allowed to hunt animals, and she wasn't really interested in fetching balls, so I taught her a version of 'hide and seek' where I'd put her on a sit-stay somewhere where I could disappear out of view, then I'd go and hide one of her fluffy toys (how and where I hid the toys grew in complexity as time went on). The key that I'd protected the classical conditioning of the cue was that I could put her on a sit-stay, walk out of sight with her toy, but her eye's didn't light up or body tense in anticipation until I said 'where is it?'. Another reason that protecting the cue is important is that if she were around children that had fluffy toys, she was less likely to tense up in anticipation unless she heard my cue, but if she did she was also aware and amenable to my direction to leave off because her hunting instincts were regularly satisfied.
The next thing is that you want the cue to a behaviour to be something that you can control and is explicitly reserved for the behaviour in question. Here's an example that appeared on my newsfeed this morning:
"A common mistake when teaching/training using food.
**********************************************
How many that are training their pups with treats are using the opening of a plastic bag or container as a cue that a treat is available?
I see this mistake so often. The pups owner decides to do a little training session with their pup. They go to the kitchen and open a plastic bag or container. Or open a container just before they start their training session. The pup comes running, 100% focused..... Training session over, close container or bag, and put it away. Most pups will very quickly switch on to this routine.
When using food to train, only the behaviour/response you are teaching should be marked with a sound or word. Don't mistakenly telegraph to your pup (or dog) that food is now available. If you are using treats, you should carry the treats with you, and randomly mark and reward responses when your pup is least expecting it. Don't go to the kitchen and open a container that cues the dog whats coming next, or open a container just before you give a command. This is one reason many that are training pups with food cannot get their pups attention unless the pup is triggered by a sound that food is now available. Carry food around with you in a treat pouch, so your pup never knows when the treat is available; not until a required behaviour or response is marked with a sound or word. Too many are unintentionally cuing the pup before the required behaviour, with a sound that food is coming. - Mark Singer, Dog Behaviourist"
The problem with the opening of cans, jars or packets cuing a dog that a treat is coming is obvious - they can start harassing people for food, or running into the kitchen when you're trying to cook a meal creating tripping or spilling hazards. So apart from Marks suggesting of keeping a treat pouch on you in the early days of training, you could also teach a training cue. I used the words 'are you ready?' for my dogs training cue. Using this meant that I could engage my dog in training without having treats or food on my person and that I could go to a packet of food to get a treat for her - in other words, the presence of food or the sound of packets or jars was NOT a reliable cue that she would get food. The most reliable thing that predicted reinforcement for her was my words.
Here is a 'dancing with dogs' routine being trained by back chaining. In this example, the clicker is a 'reward marker'. It both signals to the dog that it has done the right thing, and lets the dog know that reinforcement is on it's way.
Here is a complete dancing with dogs routine - probably one of the most famous:
Dogs are just as happy to be given tasks to do as a member of their human pack and I think it's good and healthy for them so long as the task is within their abilities, level of health, natural energy levels and resources.