I mentioned “The Model-Rival Method” earlier when talking about training dogs to associate words with objects, and I thought it’d be fun to illustrate what I was talking about. The video at the bottom of the post is an example of this method, famously used by Dr. Irene Pepperberg to train Alex the African Grey Parrot to label a large number of objects, materials, colors, etc. It was originated by the European scientist Todt, in contrast to the “Skinner Box” kind of training in which a parrot got a food treat from a mechanized box for vocalizing something similar to the sounds being played by loud speaker. Using that method, American behaviorists had concluded that parrots “can’t be taught language,” but Todt noted that it had little relationship to how our own children learn language. He criticized such a model as being a less than conclusive test of the cognitive abilities of parrots and suggested what he called the “Model-Rival” method. Dr. Pepperberg was intrigued by his argument, and set up training sessions using this method.
The training included 2 people and Alex, with one person being the ‘trainer’ and one being the ‘model’ as well as the ‘rival.’ I’ll describe a sample session as including Irene, Alex the parrot and a woman named Julie (I made that name up.) The term model is used to describe the technique because one person models what they want to parrot to do. In Alex’s case, the goal is to get him to say the word. For example, Irene would hold up an object and say “Julie, what’s this?” and Julie would answer “Crayon.” Irene then handed the object to Julie, who got to handle and play with it (thus, the human trainee is also a rival of Alex’s, because they got to play with the object but Alex did not.)
Last year I tried using a standard operant conditioning paradigm with Willie to teach him to label objects, putting down 2 toys and clicking and treating if he went over and nosed the ‘correct’ one. Willie not only was NOT able to make the distinction between the sounds I was making and the 2 objects, he became so stressed (presumably) about not knowing what he was supposed to do, he shut down. I quit after a few weeks because we were getting nowhere and Willie looked too miserable to continue. He seemed to want very much to get the game, but he couldn’t figure it out.
After reading about Chaser and his 1022 labels for objects (see my post on January 11th, 2011) I decided to try to teach Willie the names of objects in a more natural way. For about a week Jim and I used the word Scorch for his scorpion toy (a favorite, and still, amazingly, in perfect shape after how many years?). We’d toss it around, hold it up, saying Scorch as often as seemed possible. But after awhile I started thinking about the Model-Rival system, and have starting giving it a try. I’d love to tell you, after about 10 days of training, that Willie has it down pat, but he doesn’t, at least not once we added in a second name for a second toy. After about 5 days he would reliably go get the scorpion when we said Scorch 100% of the time, even if there were several other toys around , but once we added in a second toy his response fell apart. That’s not surprising, and that’s what we’re working on now, understanding that Scorch means scorpion, but Pony means the stuffed pony. You can see where we are in the video below (in which Willie boy is extremely distracted by a parked car outside and the camera especially).
By the way, out of curiosity I checked online and there is a published paper that compares the model-rival method and operant conditioning method for training domestic dogs that came out in 2003 in Applied Animal Behavior Science (by McKinley and Young). There are some pretty significant issues with the methods: they ask the dog to label any of 3 red objects as “socks” and any of the 3 yellow objects as “cross.” I’d then argue that the test is not about labeling an ‘object’ but a color (and red is not a color dogs see well). That’s a very different exercise than matching a word to an object, but granted their intent was to compare the efficacy of training methods, not so much what was being trained. Their conclusion was that the model-rival method was as effective as operant conditioning, and their conclusion is that this is a method that might be used more in dog training. The sample was small and I think there are some confounding factors, but still, very interesting stuff.
Note: You’ll notice that once I introduce Pony I say this is a “test” but then switch to saying this is “training.” I’m trying to use the same kind of method we’d use with a child, so once Willie doesn’t do what is expected (and doesn’t ‘pass the test’), I help him to do the right thing. And I’ll talk later about another issue, the potential that I, or anyone, could be inadvertently cuing their dog to the correct object…., but for now, my objective was to illustrate Willie’s progress and demonstrate the Model-Rival Method.
Dawn says
Model/Rival works well for parrots. I have one who refused to step up. Until we did some model/rival training with another parrot who would step up and get the reward(sunflower seeds). I tried a similar thing with the corgis trying to get my non retrieving corgi to hold the dumbell. I got as far as her mouthing it, but apparently I wasnt offering high enough value rewards because she pretty much didn’t want to play that game. I never tried it though with another human as the rival. Might give that a try.
eardog says
patricia…
first, i have to say that willie is so beautiful that i get distracted just looking at him. (in the past, other videos with willie were shot from a bit farther away, when he was in motion!) i had to watch your video twice to concentrate on what you were saying!
but i love this video…because it demonstrates your gentle way with him. you have such an ease with willie and point out to us how he’s interpreting the scene (it is your books/blog that have taught me to look for signs of anxiety, licking the lips and whatnot) and i found it particularly instructive when you mention towards the end that it’s perfectly fine and expected to take a few steps back in the learning process.
i’ll take your cues and continue working with my theodore (a border mutt) on differentiating the names of his toys. it’s fun work!
kecks says
I just have to say that i love the way you are working with Willie. It’s awesome. He enjoys it, you seem to enjoy it a lot, too and it’s just great to watch. Thanks for sharing!
Susan Mann says
Hi Trish – this is GREAT STUFF! I’m curious, guessing that in a prior training session you did the same thing with the pony that you did with Scorch at the beginning, modeling the correct response? So he would have had a percentage of correct responses doing it that waybefore “testing” him by asking for it without the “rival” having done it first?
I’ve been very interested in canine cognition, labelling, etc for a while, and have attended a couple of Kathy Sdao’s seminars where we worked on having the dog identify a couple of objects by name (cue) using operant conditioning, paying a lot of attention to not giving body language cues, using dark sunglasses to obscure unintentional eye flickers that might cue the dog, etc. Ideally, the dog would perform the task with the handler out of sight so as not to influence the dog, most of whom are incredibly adept at reading tiny body signals. In your demo, you used the object to reinforce the dog, and I’m forgetting (read Irene’s books a while ago, but don’t have them handy) if that is an essential part of the method, or could the animal identify something and then be reinforced by something else (a treat)? I’m thinking that it becomes harder to tease out and test whether the dog truly “knows” the word/label/cue if the handler needs to use the object to reinforce, vs perhaps having a manners minder in the room with mirrors or something so we can see that the dog did the correct behavior. Or perhaps I’m too used to OC thinking instead of RM!
trisha says
Good question Susan, and yes, an integral part of the MR method is that the object itself is the reinforcement. The way to get around the unintentional cuing problem is to separate out training versus testing… the testing could (and really should if one is doing an experiment) be done such that the dog has no visual connection with another who ‘knows’ the correct answer, but the point of the model-rival method is to link the word with the actual object in a realistic and meaningful way, thus the object itself becomes the reinforcement.
Mikki says
Similar to you, in the past I tried to get my dog to learn the names of he’s toys using the operant method but he too got really anxious and did not seem to have fun anymore so I gave up. This looks interesting and I’m rather tempted to try again but I can’t really get anyone else to help me out :/ I wonder if this could work with just one person?
Alexandra says
Very interesting! I look forward to updates. Poor Willie had sad, sad “I don’t like the camera” ears!
Susannah O'Hanlon says
Great demo, I am very interested in this method. But what do you think about the follow up paper by Cracknell et al, which suggests that actually it is just local stimulus enhancement? I haven’t read up on Chaser, but how did he learn discrimination between objects.
Amy in Indiana says
Really interesting stuff. My Cattle Dog mix doesn’t understand any nouns at all, only verbs. He knows his verbs well, though. I have never tried teaching him nouns. Maybe I will try it sometime.
I’ll be thirstily waiting for more on this subject…
barrie says
Fascinating! I would have, if I wanted the dog to know that he was on the right track with ‘pony’, praised and encouraged when he clearly looked at ‘pony’ or if I was doing true M/R as the rival, stepped in and taken ‘pony’ myself when he did not. It seems here the ‘scorch’ test was M/R but the ‘pony’ test was not? With the bc in that awesome video did they add any new toys in before the dog was 100% on each toy without another toy there?
Melissa says
Interesting. For my PhD project I’m training dogs to discriminate between sounds – one that means a reward and one that means water. We are just finishing up a pilot study and most of the dogs that got through the first phase (touching the apparatus for a reward) eventually got through the discrimination, with a couple that didn’t as they got too frustrated. Very small numbers, though. My approach has been operant without any human interaction because that’s the way it’s been done with rats, chickens and starlings, and close to in dogs and sheep. I does seem to take a long time for dogs to learn the discrimination, though. The fastest yet was a Greyhound that did it in 10 15-minute training sessions over 5 days.
Beth says
Love watching you work with Willie, and it makes me happy to see you give commands in your normal voice. One thing I notice in training class is so many people put on a stern “command” voice when cuing their dogs (even many trainers) and I rarely do that. I talk to my dogs in a normal voice and found it tends to result in a calmer dog who actually responds quicker. And who can blame them; I don’t respond well when people scold me, either!
One thing I found that helped with Jack with training the names of new toys is that if he didn’t get it, I would get it and throw it for him while saying the name in a happy, excited voice a few times. Like you, we started with a favorite high-value toy that he just liked to carry around, so bringing it was self-reinforcing. When switching to the second toy, if he brought the first (keys, in this case) I would say “No, those are KEYS, get TENNIS BALL” and then I would throw tennis ball and sing out “Tennis ball, tennis ball, get tennis ball!!!”
Now whenever we introduce a new toy I use this method: I unwrap it, say the name once or twice, then throw it for him while happily calling out the name of the new toy. Now that he has the concept down, just that one introduction is all it takes.
Since he loves to chase things, that was very reinforcing for him. The other part, of course, is that there is no game if he brings the wrong toy. What this has resulted in is if I ask for an unfavored toy, he’ll wander over and push it with his nose or bring it, then look at me as if to say “Yuck, don’t like that one” and then go off and get the toy he really wants to play with.
And I wanted to agree with the others who said what a handsome boy Willie is. I just love border collies but we don’t have the lifestyle for them, which is why I got Corgis: Herding Dog Lite.
Beth says
Susan, I wanted to add that how I got around the problem of whether or not I was unintentionally signaling my dog, ala “Clever Hans”, was to ask for toys when I myself don’t know where they might be. Since toys often go missing under tables or behind doors, this is easy enough. Jack will search and search til he finds the right one.
I can tell by his footsteps when he has it. He wanders around a bit heavily while he’s looking, then when he finds it he trots lightly and eagerly. If he’s in the kitchen (out of sight) and I’m in the living room, I instantly know when he’s found the correct toy because I hear him put on his happy gait.
On an unrelated note, what I find amusing is if Jack’s having trouble finding a toy and my husband starts looking, the dog INSTANTLY stops searching and starts following my husband. Another clue, as if we needed it, that to many dogs– even clever ones— their default problem-solving strategy is “Humans are smart and have thumbs; let them figure it out.”
However, in this case the dog is infinitely better than the people at finding something, so I have instructed my husband to not get involved in the search. If something shows up where the dog can’t reach it, he’ll bark. That is probably part of the reason why tracking dogs are worked on a long line; if they were close to the person, they might default to watching their human instead.
Frances says
Very helpful – not least because you emphasise that you have been training over weeks, not just minutes or hours (now, where was that post about how long it takes us to establish a habit ….!). Like others on here, I am one human with two dogs (and two cats – one of which responds very well to reward based training). Have you any tips on how to use MR training in this environment, without setting my animals up for squabbles? The three of them are getting pretty good at sitting in a row for treats, and waiting their turn politely, but that is about as far as it goes.
Also I have one dog – my papillon Sophy – who likes to get things right, and is easily switched off if she doesn’t understand (perhaps I was too quick to solve problems for her when she was a pup). And the other – Poppy the toy poodle – who flings herself into every game with huge enthusiasm, but doesn’t always stop to think. A future post on working with the different learning styles dogs have would be enormously helpful!
jackied says
Lovely video!
A lot of ordinary people must use the names of toys and reinforce the dog by playing with the toy and talking about it in quite a natural way as they do so. We certainly do. It’s the same way as humans talk to babies and toddlers. But by the sound of it most dogs still don’t learn the names of the toys.
Michelle says
Great explanation and video! I too love watching you work with Willie, your methods are wonderful.
I have a small collie mix who can go and get one of about 10 different toys. Though we never formally taught this, I think we inadvertently did use the MR method. I would hold a toy and say in an excited voice “Oh look at the (bear, for example)!” Then pass it to my husband who would do the same, then we’d throw it, always repeating the name. Then when she got it we’d say things like “you got the bear!” or if she brought it back “thank you for the bear!” We didn’t do this to teach her the names of objects, it was just us being enthusiastic! But after a few weeks/months, she definitely knew the names of her toys.
Jen Gibson says
Does Pony get the same kind of excitement level from Willie when you just play (or he just plays) with it? With the camera off? 🙂 I’m just curious. Because if Scorch has been the toy that has given him the most fun, the most time with you, then that would always be his choice. And if Pony is only the other one that comes out to see if he knows the difference (and therefore means less time with Scorch) then maybe it’s more about choosing which one brings better or more play time with mom?
It’s all very interesting! If only they had little balloons above their heads with their thoughts in them! 🙂 Thank you for sharing.
trisha says
Argh, just wrote a comment responding to your comments, and it got wiped out by random computer evil forces. Sigh. Gotta get to campus, but here’s a shorter version, with apologies for being abrupt…
Thanks Beth for sharing your story and reminding me that I too said “No, that’s Pony..” once or twice to super sensitive Willie, and set him back about 2 days. I think he interpreted it as “Don’t do that” and who knows what ‘that’ was but he wouldn’t pick anything up at first. I went backwards and just played with Scorch for awhile while saying its name.
Comments to other great comments and questions: Willie loves Pony as much as Scorch as best I can tell, yes I mixed M-R with plain old reinforcement in the video, good point. I should be clear that is what I’m doing every day, mixing the 2. If someone else is there we do M-R, if not we play with the toy, I say it’s name a lot and we try to have a great time. Willie gets reinforced for going to the toy and picking it up. And no, I did not give Willie as much ‘training/conditioning’ with Pony as I did with Scorch.. I’m thinking I went too fast. So right now we’re just playing with Pony. I have put Scorch up for several days….
Would love to write more but gotta run to UW and teach my class!
Kat says
This is fascinating reading and seeing what you’re doing and following the comments. I’m starting to think that perhaps just as some humans have an easier time with math than others and others find the verbal skills come easier dogs come in the model that’s all about doing things (the verb dogs) and dogs that find it easier to understand names for objects (the noun dogs). Ranger seems to take longer to learn actions than he does to learn names. My son is trying to teach him to carry a rope handled bucket in his mouth. After one training session Ranger knew bucket as the object he was supposed to perform an action on and will go to the bucket, look toward the bucket and generally indicate that he knows what the bucket is but he still needs a lot of coaching and luring to take the bucket handle. I know some of it is working with a new trainer and one who is young and less consistent than he should be but that didn’t stop Ranger from learning that the sound pattern “bucket” means a specific item. Clearly there will be lots more training experiments in his future which he’ll like since for Ranger there’s nothing quite as much fun as any kind of training.
Barb says
Great video and discussion, Trish. Thank you.
I was wondering how much of a difference it made that you were the model/rival. Would the fact that you stood next to Scorch be important? Could you tape the training session if you stood next to Willie across the room from the toys?
Don’t have good answers to these questions, but have decided to try to do some toy discrimination work with my dogs. I know in agility, dogs can discriminate between objects based on the vwrbal name of the object in a distractionless setting, for example a table and a tunnel. I also know that in the heat of a “run,” a dog will go into a tunnel that we are running toward even if I call the dog a table. Hmmm.
Josh says
Thanks for all the recent posts. I find the animal/language fascinating on so many levels. The confounds on teaching another species (or even the same species at different developmental stages) are so huge. Are we just “talking” past each other?
Did you seen this.
http://www.npr.org/2011/01/20/132650631/new-language-discovered-prairiedogese
So these Prairie dogs know nouns (hawk, human) but also adjectives (“tall”, “blue-shirted” human). I would venture a guess that they also know verbs (blue-shirted human “coming,” short human “leaving”).
It all seems so anthropomorphic to me. We define language and the teaching of language as being something “we” understand (and/or can perceive). But just because we can’t hear, see, smell “it” (because we are limited by our physiology) doesn’t mean it doesn’t exist.
Josh
Beth says
I just wanted to add that it seems some dogs are way, way more cued into the subtelties of human language than others, and those who are cued in can be a blessing and a curse.
It seems that some breeds, and then individuals within breeds, are much more laser-focused on figuring US out than others. My language-oriented dog is also the one who gets upset if I’m mad at something, even if I don’t say a word. He’s the one who comes and lays on top of me if I’m sick (even though he’s not a cuddler). He’s the one who gets upset if we do something different and he can’t figure out what we’re up to. He’s the one who comes running to “help” if someone says “ouch” in even a normal tone. And he’s the one who somehow, on his own, learned that “excuse me” means “move out of the way.”
My other dog does not learn words nearly as quickly, but is also not so likely to get upset by small changes in OUR behavior. The sensitive verbal dogs are great fun but wow, you need to be careful with them and I’m not sure they are appropriate for more chaotic homes (even if it’s good chaos).
Beth says
I also find it fascinating that Willie gets upset when you tell him no, that’s not it. Jack is also super-sensitive (no harsh corrections for him). However, he seems to appreciate the directness of “no that’s not it” when he’s offering behavior that is not what I’m looking for.
Just like people, I suppose some dogs are mortified by any corrections, while others like the up-front approach that tells them “that’s just not working; try something else.” Just goes to show that there is no one-size-fits-all approach to training!
em says
So interesting! I’m terribly impressed with all the dogs who learn to discriminate between nouns, particularly because Otis seems to find it so difficult. He’s another one who hates being “wrong”, and becomes too stressed to make all but the gentlest training worthwhile. Weirdly, he doesn’t seem to get as stressed about corrections. He knows the ‘no-no’ tone and seems to take it in stride, stopping his behavior without any fuss, but mixed or confusing signals send him into a spiral of mental anguish that takes hours to recover from. Almost all of the precious few words that he knows (his name is the one exception) he picked up accidentally by making an association with a positive outcome (food, daddy, grandma).
Amy W. says
Has anyone tried the M-R method to teach behaviors like sit or come? Was it effective?
Frances says
Amy W – I’m not sure if it counts as MR, but my two dogs know that if they want me to fill up the Brick Toy with treats, they both have to lie down. If one is down, and the other not, I will pointedly praise the one who is in the right position, and wait for the other to catch on. I don’t cue them for position in this context – Sophy always used Down as Please, and it grew from there.
MM says
When you first got Pony and dropped Pony and Scorch on the floor, Willie looked away. What does it mean when dogs do that?
Steve says
Several years ago now, I had a Britt mix. I never really thought about his “vocabulary” until I read “The Intelligence of Dogs” by Stanley Coren. It got me thinking about the a actual process. This may be a way to work with Willie.
Pretty simple, but when Talbot would bring me something, I would take it and throw it for him, then say “Get your X.” So, for a tennis ball, for example, I would say “Get your tennis,” for a nylabone, I would say “Get your toy,” or “Get your Kong” for the Kong, etc. Once he associated the word with the object I would then start request specific things. So, if we were sitting in my apartment, I didn’t want to throw something hard like a nylabone, so I’d say “get your tennis” which he would then differentiate from the other objects he had scattered around the apartment.
By the time Talbot passed, aside from basic obedience cues, he could differentiate between about two dozen different objects. (I have to admit I was kind of deflated when I read about 1022 words Chaser know.)
Lacey H says
I’m not much of a trainer at all, but my foster dogs generally learn “sit” by MR, with the other dogs being models and rivals. Most are fluent in “sit” after a couple of weeks.
Kathy Haig says
Lacey’s post reminded me: We owned two dogs who were very well-trained (thanks Dogs Best Friend!) and then we began fostering a rescue dog who knew NOTHING and without knowing it was model/rival (I had never heard of Alex the parrot at that point, over a decade ago), I used that technique to teach Baern, the rescue dog, the basics.
I brought them all into the kitchen (a common training area in our home because training started when they were puppies with waiting to eat) and lined them up, asking Loki and Rainy (the house dogs) to sit. They sat and got treats. Then I asked Baern to sit. Nothing. So, turning back to Loki and Rainy, I had them stand and then said “sit” again. They sat and got treated. I asked Baern to sit. Nothing. This happened only three times before Baern’s back end sank slooooooooowly toward the ground. JACKPOT! Joyful jumping around and throwing treats like it was a parade!
Until his hips got old and sore, Baern sat every time after that without ever forgetting it. He was an extraordinarily fast learner–especially whenever there were treats involved–and watching the other two dogs get treats while he went without was really frustrating for him. I think the behavior he offered (sitting) was spurred by that frustration. It seems like a fairly straightforward idea: “they do X when she says Y, and they get treats. Huh.”
So far I haven’t used model/rival with Mico, my 9 month old, because Baern is too deaf and blind to pay much attention to any commands any more, so he’s not a very satisfactory model, but Mico is a competitive little bugger and very smart, so I think it could be very successful. Perhaps I could get my husband to play the model/rival.
Thanks for the ideas, Trisha and other commentators!
Marguerite says
Frances & Kathy, you’ve answered the question I had in mind.I have a neutered male rat terrier maybe 6 years old and later got a skittish female rattie now about 3. I have them teaching each other behaviors all the time–the boy knew how to sit up and beg (cue: “please”), and the girl quickly learned it. She can jump straight up (cue: “boing!”) and he learned it very quickly. They’re highly food motivated, but she’s a bit “softer” than he is.
I’ve moved from out in the country to a little house in town and I’m using a lot of string cheese to distract them from barking out the windows and every cat, squirrel, and postal carrier.
Lolly says
Josh, I agree with you. I think this is a facinating topic, and I also think that we want to find a way to make our dogs understand language (our language) irrespective of what they are good at in terms of communication. I think Trisha has been a fabulous teacher about how to learn “their,” language.
I have tried, unsuccessfully, to teach my dogs names for their toys over the years, but haven’t really focused on it intensely.
I like this Rival Model.
I am also grateful to see someone whom I respect encouraging people to use other models than OC. I love clicker training, and think it’s been a blessing for me and my dog, but I’ve felt that it isn’t the only way, and have been waiting for other options to teach behaviors than just the dominence model and the OC model.
Jessica says
I’ve taught my dog to touch either a red or blue frisbee when I ask for it (about 95% reliably) and as an experiment, switched to two books, one very red and the other very blue, and asked for the same behavior. My girl was able to switch from picking a blue frisbee to a blue book almost immediately (one error in the very beginning). Do you think this argues for an understanding of the concept of “blueness”?
Donna Wolff says
I casually began overlaying a name whenever my dog was playing with a particular toy and using it to interact with me. Not knowing about Chaser, thought my Gisele (giant schnauzer) was a genius as she learned the names of each of her toys and would bring when asked. Still think she is a genius….just not a border collie and I am not nearly as motivated as the gentleman with Chaser.
.
She can discriminate between green bone and rubber bone, identical except in size and color (I dont believe she knows color, just given name), knot bone, fist bone and Dino (all nylabones) and her many real bones are just Bone. She will select between chuck-it ball, whistle ball, jingle ball, grass ball (and teaser ball which she knows must stay outdoors).
We usually have a limited number in her box (6-8); I have never tested her on all thirty or so, however do impress dinner guests by asking her to bring whichever are presently in the room. When I rotate toys from closet to her box…keeps them more exciting and the floors less crowded, I find we need a short review.
When replacing the Tuffy stuffies like the bull and vulture, I was concerned about using the same name (ours, not the one on label) and began adding two after their names (Tex Two,, Vern Two), however soon dropped it as I forgot and it seems she did also! I do let her chew these and she will remove small pieces of the edging which she loves to bring and we play Open (mouth)I place fingers inside and play Where is it? find it then toss and tug on1-2 inch pieces! She loves this! 3yo and has never swallowed one.
What I think I saw: 1) The camera and operator took what is a very offhand part of daily play in my home, created an emotionally stilted environment and as you noted, made Willie nervous, never the best learning mindset. 2) During the last trial Willie sniffed Scorch, rejected him and moved to stand by Pony which I would have marked as recognition even though he did not pick it up. 3) Willie appears exceptionally sweet and well mannered, not the temperament I think would challenge a rival….especially after watching you get so excited about Pony could he have thought …hmm she likes Pony so much she can have him…I like Scorch better anyway.
JW says
Fascinating! Seems the dog distinguishes between the two objects and chooses the one that he gets all the praise/attention for choosing. Was this the goal?
Doesn’t show that the dog understands the name of the object or what is being requested of them. Dog has learned, if I choose object A (bug)I get praise, verses if I choose object B gets no praise.
Interesting…
The initial training steps when the human picks up the object reminds me of sibling rivalry.