Training in the Multi-Animal Household

Melissa Alexander's picture
Filed in - Training - Dogs - Clicker Digest

I've seen this topic come up on multiple mailing lists recently: "I have two <insert favorite animal>. Do I need to separate them, or can I train them at the same time?" My standard answer used to be "Separate them," and I still believe that's best when shaping or when you (or your animal) really need to concentrate. But there are benefits to training in groups.

First, the animals learn from each other. From sled dogs to hunting dogs to carriage horses, trainers have paired experienced animals with inexperienced animals for centuries, letting the older, experienced animals teach lessons in a far more efficient, effective way than the humans could. Although scientists dispute that some species, such as dogs, are incapable of observational learning, anecodtes about animals learning cues and behaviors from watching their buddy's training sessions are common. One could argue whether the learning is observational or social facilitation or allelomimetic, but the bottom line is that training in groups or even just letting an untrained animal observe can speed up training.

I learned this when my curly coated retriever was young. I was having a hard time teaching him to back up. My Newf butted into the training session and responded to the cue to back up. Pax watched, offered the same thing, and hasn't had trouble with that behavior since. With my youngest dog, Aslan, I have intentionally done most of his basic training with Pax. Rather than have sessions where I work with Aslan on sit, down, stay, and come, I just paired him with Pax and cued those behaviors. Aslan copied Pax to try to get the reward... and learned the cues along the way. It was fast, effective, and fun!

Another benefit to working in groups is the spirit of competition can result in faster performances. Want to speed up your recall? Call the group and reward only the first to arrive. Want to speed up the sit? Cue the sit and reward only the first to sit. Working in groups can also help you teach a dog to respond to cues preceded by his name and only cues preceded by his name.

When you or your dog needs to focus, by all means, work without the other animals present. But if you have more than one animal, try working in groups and see what benefit you can derive!

About the author Melissa Alexander is a writer and clicker trainer in Seattle, WA. She owns the highly popular ClickerSolutions mailing list and is the author of Click for Joy, the award-winning, essential guide to clicker training.

Group Training & No Reward Markers

In my experience dogs do benefit from being trained together at many points in their training. The only problem I have had concerns using NRMs. My lofty goal is to never use them, but the reality is that I do use them. Some dogs take it well, as I intend it. Just as a cue that they need to try something new. Other dogs do not take it too well. Sometimes I know I have made a mistake in my training, some dogs are just more sensitive, and training other people's dogs I am dealing with a lot of baggage at times where a dog has had negative training experiences. I have a rescued Golden Retriever who came with some of that baggage and is very sensitive as a whole, so I need to be very careful with him. When training dogs together I have learned from Max, my Golden, that a NRM meant for one dog can have a very negative effect on the other dog. Worse, the dog suffering from an undeserved NRM is usually behaving very well so the NRM is telling him not to do something he should actually be doing. If you use NRMs be careful when training dogs together so that one dog who is doing well isn't taking a NRM to heart that was directed at another dog.

Jenny Ruth Yasi's picture

NRM

I think the trick is to make sure that the NRM is NOT associated with punishment, but rather that it is neutral information the dog can USE to find reinforcement. I go way out of my way to make sure that NRMs are as reinforcing as possible. That maybe sounds weird, but what it means is that my management tools (leash, tether) are reinforcing places. The NRM, which does precede a puppy going onto a tether because she's been harassing the kitty) is as minimally as possible associated with punishment, but is maximally as possible associated with information that helps the dog find the reinforcment.

Animals can get very frustrated when they are working hard but not paid. So, when a trial "fails" (IE: I offer a cue but the dog does not respond appropriately) instead of a NRM, I often offer a "release" cue ("okay"). The release cue says "that trial is over" and the release cue is often, but not always inherently reinforcing. So I might say "okay" as I release the dogs (I don't need you to wait anymore, you can go and play in the yard), or I might say "okay" (I don't need you to find the frisbee after all).

I discovered this useful application of the release cue, because if I offered a "sit" cue to my bichon, and he didn't feel like sitting, he learned to wait and see what other cues I might have available. He'd pick and choose which cue he wanted to respond to. So, rather than discouraging him with a whole lot of "oopsies" and pschological paw twisting, I learned to give him a one shot deal. "Here's your sit cue, take it or leave it. You don't want the sit cue? "Okay" This made cues a much more special, limited time offer, for him.

I think it is MUCH better to have a sightly ambiguous cue, IE one associated with both reinforcement and no reinforcement, than to have a cue associated with aversives. I am very clear that punishment doesn't simply punish behavior, it punishes the whole relationship.

Dogs are familiar with interpreting slightly ambiguous cues. It is a healthy exercise for them to figure out what "oops" means. Maybe it means I just tripped on my own shoelace; maybe it means (if I have cued "find Daddy") that Albert isn't upstairs. If "oops" (for example) is sometimes associated with reinforcement (the dog uses the oops to find Albert more quickly) and with no reinforcement (I stop and tie my shoelace), this helps the dog understand the NRM as raw information. Information is reinforcing, because the dog can *potentially* use it.

I agree 1000% with you that the NRM can be toxic and one must be very careful to avoid discouraging rather than encouraging ourselves and our friends. Sometimes, easier said than done!!

Yasi

BlueHairBob's picture

Multiple parrot training

I currently have two parrots - and training them in the same room appears to be a great way to go. I work with one at a time right now, but the other (and the dog) are always watching. The parrots seemed to be so intelligent that they pick up the training in some cases just from watching. Having the other animals in the room does cause more distractions, but the benefit of all of the animals learning in the same session may outweigh them. I think that they even learn what to eat and how to be handled by watch each other. Since this is how they learn from each other in the wild, it only makes sense that it would enhance our training.

Jenny Ruth Yasi's picture

One click, two treats

I train both my dogs separately, and together.

When I am working them together, if I click, I treat BOTH of them, because otherwise I am concerned about degrading the click. I don't want Dandylion to hear me clicking while I am working with Tigerlily right next to him and yet somehow Dandylion doesn't get a treat. That's sort of heartbreaking, it's not fair, and my dogs don't like it, so I try to avoid it. When I know my training focus is on Tigerlily (for example) I make it really clear by either tethering or crating or putting Dandylion in another space.

So when working side by side with the two dogs, it is one click, two treats. Click criteria? They must BOTH be doing something right. Often, it appears I am clicking and treating one dog simply for not interrupting the other, or for discriminating the difference between his cue and the other dog's cue. If I ask for both dogs to sit, but only one dog promptly sits, I might give that good sitter a little bit of food, but I save my click for those times that BOTH dogs sit promptly when cued to do so.

I have one dog jump over the other dog. Success! So, that's one click and both dogs get a treat. If the jumping dog lands clumsily onto the jumped dog, I give bumped-into dog some food reinforcement, but I don't click.

First I do train the behaviors separately, but then we really enjoy practicing together.

Group training thoughts from a musher

Since my dogs have to work as a team I do much of my training as a team, but one-on-one time is also important. Each of my dogs is working on a different behavior (for example, a youngster working on 'four on the floor', a newly rescued team dog working on loose leash walking, &c.

I think the handler should pick and choose the setting and technique most appropriate for training the desired behavior, keeping in mind that dogs will learn inappropriate behaviors from other dogs just as quickly as they will learn desired behaviors.

Swanny

"A good dog is so much a nobler beast than an indifferent man that one sometimes gladly exchanges the society of one for that of the other." William Francis Butler

teaching touch

I have two dogs, frannie and wilson, standard poodles. I am new to clicker training. After reading the Whole Dog Journal article about teaching touch, I thought I would try. I am training them together. That doesn't seem to work. Neither dog has really gotten the idea yet, though at times I think maybe. I guess I should separate them, totally. I have tried working one dog and then the other, allowing one to watch. but even with that, the working dog just doesn't seem to get the idea that I want her or him to touch the stick.

I must be doing something wrong.

Teaching Touch to Two Dogs

As you have two untrained (at least for this behaviour) dogs I would certainly separate them while working with one at a time - the other one can be in another room or in a crate.

You don't mention whether the target you want them to touch is a target stick or your hand, or something else, but if you are not getting touches you need to start off by shaping the behaviour. This means that at first you C/T even a look in the direction of the target and work from there. Don't be afraid to maniplulate the target a little to help get those looks - I find that it is a rare dog that won't at least glance in the direction of a stick that is being moved into touch range and out again, or a hand that is comming from behind your back into touch range. If necessary, rub your target in something that your dog loves - if nothing else works the smell of a favorite treat should have nose touches happening in short order. Good Luck!