The Other Thing

Audie has learned how to match items on command.

The ability to categorize, to sort things out based on the kinds of features they share, is fundamental to  perception, learning, and judgment. The ability to sort things into various mental categories allows animals to store perceptual and conceptual information in an efficient and adaptive way. Without the ability to categorize, every object and every event would be perceived as unique and it would be impossible for animals to generalize and learn.

Boston.com reports on research being done at Harvard on canine cognition:

Scientists are also drawn to dogs because of their unique history growing up in the same environment as people, and they hope to learn whether domestication has led to dogs that think and act more like their masters – or whether we just think they have human traits.

“Here’s this species we live with. Everyone has their views about how smart they are. No doubt we are overinterpreting – and in some cases underinterpreting,” said Marc Hauser, a Harvard professor who has long studied cognition in cottontop tamarin monkeys and who heads the new lab. “To what extent is an animal that’s really been bred to be with humans capable of some of the same psychological mechanisms?”

Can dogs understand such abstract concepts as “same,” for example? Or, can dogs be patient? To answer such canine conundrums, Hauser is recruiting both purebreds and mutts and running them through simple tests. In return, they earn tasty treats.

Based on a few decades of experience I’ll vouch for the fact that dogs (like kids) can most certainly learn to be patient.  I’m also convinced now that dogs – at least some of them – are capable of understanding the concept of sameness.  In fact, for a while now I’ve been working on teaching young Audie to demonstrate that he can do it.

The video below is a brief demonstration of his skills in this area.

*&%$ sound didn’t come though. Oh well.

The only hep I give him is the cue ‘other one’ and praise and petting when he comes back with the correct item. The first item is a roll of purple vetwrap; the second one is a plastic kennel cup; the third is a work glove; the fourth a plastic bottle full of water; and the fifth one – when he starts to get bored and needs a bit of help – is a metal spoon. Bad trainer. I should have used that in the second or third rep because it’s not his favorite thing to pick up. The big correction he gets for making a mistake that last time is me laughing and calling him a goon, then telling him to try again. He gets it right on the second try.

As you can see, Audie is consistently able to correctly identify which item in a small group is the ‘same’ as the one I’m holding.

I couldn’t find detailed information on the studies being done at Harvard but based on the blurb posted at Boston.com it appears that researchers are taking green dogs and testing them to see if they naturally and intuitively grasp the idea of  sameness as it applies to how abstract symbols and photos can represent real world objects.  If this is really how they’re going about it then I think they need a sharp smack on the bottom with a newspaper.

Given the fact that human beings spend years of time teaching basic concepts like ‘same’ and ‘different’ to our own children, it makes no sense to expect a naïve dog to understand abstract symbolism at the first go.

I used shoes to introduce Audie to the idea of sameness. The OddMan has a thing for shoes. He loves to carry them around the house and has a rather inconvenient habit of leaving them in odd places. Shoes come in pairs so I started by showing him a pair of matching shoes, handing one to him to hold (i.e. fetch), then taking it from him, pointing to the matching shoe and telling him ‘fetch, get the other one’. The main tools I used were a trained retrieve, directional cues and overlaying.

After showing Audie this just a few times he seemed to grasp the idea that when I held up a shoe and said “other one” I wanted him to pick up the matching one and hand it to me.

From there we added distance to the game. Instead of asking him to hand me a shoe right at my feet, I put the matching one a few feet away. Bit by bit I increased distance – then we added difficulty. Starting with them up close, I put two different shoes next to each other and asked him to get ‘the other one.’ I had to coach him a bit at first, but he picked up this idea pretty quickly too. Once he did, we put distance and difficulty together – and I had a dog who would go find me the matching shoe I wanted on command. Gloves and slippers were an easy step from there.

It took a bit longer to teach him that the concept also applied items like tools, water bottles, metal spoons etc., but as you can see in the video, he certainly appears to understand the idea now. Audie still isn’t very good at matching items when he’s distracted, and he seems to get bored with the exercise fairly quickly (typically after 3 to 5 repetitions).

Instead of expecting a young, naïve dog to intuitively grasp the idea of ‘sameness’ I used a step by step process to teach  him what it meant. And I think that I got pretty amazing results.

Clever Maggie

I came across an interesting clip on YouTube today featuring a very clever little dog called Maggie.

A bit of googling turned up an article in the Mountain Xpress News that tells us Maggie’s story:

When Jesse and Arthur Treff got Maggie, a Jack Russell terrier as a puppy, “she was a challenge,” explains Jesse. “Between chewing and just everything a puppy can do, magnify that by 10 — that’s what a Jack Russell puppy does.”

In an attempt to curb Maggie’s misbehavior, the Treffs thought they’d try redirecting her energy into learning tricks. It was immediately apparent that she was a natural. “She was picking up new things every three or four minutes,” says Jesse. “She’d get it. I’d come back and she’d still have it. It just accelerated from there.”

[….]

When Maggie was about 6 months old, a friend of Jesse’s said to her jokingly, “I bet she can even count.”

Not really expecting anything, Jesse held up four of her fingers for Maggie — who proceeded to tap her right front paw on the ground four times. Over the next couple of weeks, Jesse continued showing Maggie varying numbers of her fingers. “She was getting it right about 80 percent of the time,” says Jesse.

[…]

Initially, Maggie would respond only to Jesse, but eventually she began to answer mathematical equations from strangers as well. Sometimes, Jesse says, she still can’t quite believe that Maggie is able to do what she does. “I think she’s not going to be able to do it one day — it’ll go away. Someone will ask her something and she’ll just stand there.”

This summer, WLOS-TV did a short segment about Maggie’s ability to count and the piece was picked up by other stations nationwide. Since then, says Jesse, “I can’t go down the sidewalk in Asheville without someone saying, ‘My friend doesn’t believe your dog can count — can you show him?”

[…]

When asked about those skeptics who believe Maggie’s counting is some kind of hoax, Jesse says: “In Asheville, there are very few people who don’t believe Maggie can count. The few people who don’t believe it … I use to work hard at trying to change their minds. Now I feel like, ‘That’s fine. I couldn’t care less.’  And I really understand: You have to see it to believe it.”

Ah…  but seeing isn’t necessarily the same as believing – or more importantly, the same as knowing. 

Should we believe reports that dogs can do mathematical calculations?  Is Maggie’s owner telling us the truth?

Well… in science, as in all things, there’s a big difference between honesty and reliability. An honest observer can be unreliable and a reliable observer can be dishonest. Honest observers can make mistakes by focusing on the wrong aspects of an experiment or they can be fooled by their pre-conceived notions.  Either way, their mistakes can be gosh-darned hard to ferret out. 

Human beings have an odd tendency to be most impressed by animals when they seem to be able to mimic human activities like doing math or using language.  We are generally much less impressed when they excel at tasks that are beyond our abilities.  Obsessed with our own big brains, we also have a hard time accepting the idea that we aren’t always in conscious control of our actions.  Because of this we sometimes prefer to “dishonestly” believe that behaviors that occur outside our conscious control are governed by supernatural forces like mind reading or telepathy.

A classic example of this phenomenon is illustrated by the story of another clever animal – Clever Hans.  Hans was a horse who learned to cleverly and correctly respond to a range of questions involving mathematical calculations and other advanced cognitive tasks by tapping his hoof.  Hans was a sensation.  People flocked to see the horse that could think like a man.

Or could he?  In 1904 Oskar Pfungst discovered that Hans wasn’t spelling, doing mathematical calculations or telling time; he was responding to incredibly subtle physical cues he picked up from his owner, Wilhelm von Osten.  How did Pfungst solve the riddle that so many others had missed?  Well… everyone other than Pfungst who watched and studied Hans focused on the horse’s performance.  Pfungst was the only one who asked von Osten how he had taught Hans to tap. He was also the first investigator who didn’t automatically assume that the key issue in understanding the phenomenon was how Hans learned the answers to the questions. Because he thought outside the box, Pfungst was able to demonstrate that Hans was only able to perform well when certain people were present.

Robert Yerkes later noted another important issue with regard to reports of clever animals:  The fact that they regularly responded correctly to questions about things they had absolutely no previous knowledge of.  Clever Hans would spell out a word that included letters he hadn’t been taught to tap out.  A clever dog called Roger appeared to spontaneously learn to do complex tasks like multiplication.  These animals seemed to acquire advanced skills by osmosis.

The skills of clever animals like Hans, Roger and Maggie are astonishing.  They’re just not amazing in quite the way most people assume they are.  Clever animals don’t understand spelling or  mathematics – they’re geniuses at observing and responding to human behavior.

How do they do it?

Mirror neurons may be the key.

Horses and dogs are social species. Being able to understand the intentions of others is a fundamental part of social behavior. While the neural and functional mechanisms behind social intentionality are still not well understood, a recent article by Iacoboni, Molnar-Szakacs, Gallese, Buccino, Mazziotta, and Rizzolatti  in PLoS Biology holds some tantalizing clues.

Recently, the discovery of a special class of neurons in the primate premotor cortex has provided some clues with respect to such mechanisms. Mirror neurons are premotor neurons that fire when the monkey performs object-directed actions such as grasping, tearing, manipulating, holding, but also when the animal observes somebody else, either a conspecific or a human experimenter, performing the same class of actions. In fact, even the sound of an action in the dark activates these neurons.

Researchers have believed for some time that mirror neurons might be the neural mechanism that allows us to understand the intentions of other people. The basis of this idea arises from the fact that mirror neurons focus on actions.  An action includes both an actor and a goal.  The idea of a goal implies intention.  This action – goal – intention chain may form the basis of the process in which mirror neurons allow us to create internal representations other’s mental states in our own minds.

Context appears to play an important part in the action – goal – intention chain, and it may be the missing link that explains how clever animals are able to discern and correctly interpret astonishingly subtle cues about the intentions of the people they spend their lives with.  Iacoboni et al. propose that context provides vital cues that help us clarify intentions. They state that, “The same action done in two different contexts acquires different meanings and may reflect two different intentions.”  

Context provides the vital contrast that allows an animal to differentiate one subtle action from another.  If I call my dog’s name while I bend every so slightly and smile he is able to tell that my intention is entirely different than it is when I call him from a somewhat more formal posture.  And he doesn’t need to be able to read or do math to figure it out.

And – context is important in our discussion here.  Unlike humans, dogs and horses aren’t distracted from subtle changes in context in their sensory world by a continuous internal narrative. Free from that endless stream of white noise, they exist in a world that’s far richer in sensory input than ours. This explains why they’re often startled or distracted by sights, sounds and smells that we can’t even detect.

Given their much greater sensitivity to context and sensory information – it doesn’t require a great leap to imagine that clever dogs and horses can learn to detect and respond to incredibly subtle – and even unintentional – physiological cues emitted by human beings.

Why is this important to pet owners?  The important issue here is that of our common, human failures of expectation. We take a naïve, inexperienced animal and put it into a situation where we expect it to perform advanced tasks like an experienced, trained  human would.  If the animal fails our expectations and doesn’t respond “correctly” we assume that it’s stupid.  If it meets our expectations by responding correctly we assume that it has suddenly – and without prior training – successfully made the mental leap to being capable of advanced human skills.  And in both cases – we are the ones displaying a sobering lack of cleverness.

I run into errors of expectation nearly every day in my work as a dog trainer.  A busy family gets a new dog and they expect it to learn the rules of the household by osmosis.  When the dog, understandably, fails to respond in a properly “clever” way, they assume that he’s stupid and give up any hope of  training him.  Worse yet, if the dog initially manages to parse out a reasonably acceptable version of the rules on his own, he’s likely never to be given the training and attention his – obviously bright – mind craves because, of course “he doesn’t need it”.

We can learn far more about the mental abilities of animals when we learn to see them they way they are instead of the way we’d like them to be.  Your dog shouldn’t need to learn to read or do math to impress you. Once you learn to recognize them, I’m sure you’ll find that your dog’s got plenty of astonishing skills – and he’d love to share them with you.

So show me who’s a clever girl.  Spend a few years teaching your dog to understand the alphabet, spelling and phonics –  and I’ll be thrilled to celebrate your success.  Spend a few weeks or months teaching your dog to read subtle physiological cues in a way that makes it look like she can read – and I’ll be more than happy to celebrate her success.

Games People Play — With Dogs

The key to understanding what constitutes a good game versus a bad game to play with your dog is:
When YOU are the one guiding the play in a proactive way, it is a good game. 
When YOUR DOG guides the action and you are simply reacting to what he does, it is a bad game.

Good games are an excellent way to tire out your dog by using his mind.  They can also be an enjoyable way to proof your dog’s obedience.

Bad games are an excellent way to teach your dog that misbehaving is a way to get your attention.  They can also teach your dog that it is okay to ignore your commands.

Here are some examples of bad games and – some ideas for good games.

Bad game: Wrestling and rough play
Many people like to wrestle and play roughly with their dogs.  When we engage in these kinds of games with our dogs we encourage them to treat us like another dog.  Think about it.  When two dogs play together, they jump up on each other, bite each other, chase each other and test to see just who has more power.  This is a VERY bad game as it teaches your dog to rudely jump up on you, bite you and challenge you.  Sadly, I see many dogs that start out playing these kinds of games with their owners (often children) and end up biting them severely.

Good game: Teaching tricks
Teaching your dog to do tricks exercises his mind and his body.  There are some great books on teaching tricks available at your local library or book store.

Bad game: Catch me if you can!
Your dog grabs an item you don’t want him to have.  You call him to you to try to take it away, and instead of coming he runs off with it in his mouth.  You chase after him and try to catch him to get the item back.  This is a bad game because it teaches your dog that he can out-run and out-smart you.  It is also bad because teaches him to run away from you instead of coming when you call him.

Good game: Hide and Seek
Find a place where there are only minor distractions.  Hide in a place where you are easy to find and call your dog so that he can search and find you.  Give him a BIG reward (love, petting, play, treats) when he finds you.  As he gets better at the game, hide in more difficult places or play the game in areas where there are moderate to strong distractions (depending on your dog’s skill and reliability). This is a good game because it not only teaches your dog to come when you call him, it also builds a drive to search and find you when you’re not in sight.

Bad game: Wild or aggressive tug games
Playing a wild, aggressive, no-holds-barred game of tug-of-way with your dog, a game where he growls and bites at your hand to win, or where he learns to run off with the toy is a bad game.  This game teaches your dog that if he challenges you, he might win.  It also tends to make a dog more likely to bite and nip at hands and often turns into game where the dog runs off with the toy and ignores you.

Good game: Polite tug games
In polite tug-of-war games you initiate the play by showing the toy to the dog an inviting him to play tug with you.  Hold the toy and give the dog a command like ‘tug’ or ‘take’ and encourage him to take the toy in his mouth.  If the dog plays too roughly, you must end the game immediately and start a training session with the dog.  If he plays nicely, continue on a bit making sure that YOU decide when each bout of tug ends.  Unless you have a timid dog, only play this game in small bits. Be sure to end the game immediately when the dog gets too excited or nips at your hands.  This game builds self-confidence in a timid dog.  It can increase self control in dogs and it shows the dog you make the rules – but ONLY if you play it correctly!

Bad game: Throw for me
If your dog shoves or tosses a ball or other toy at you and TELLS you to throw it for him – you are playing HIS fetch game.  If your dog refuses to release you the toy you’ve thrown or won’t bring it back to you, he’s playing a game with you.  These games teach your dog to challenge you, to ignore your commands and to see you as a fool to toy with.

Good game: Fetch
In a good game of fetch, you initiate the play by showing the dog a ball or toy, telling him to sit and then throwing the toy after he sits.  The dog politely brings the toy back and either drops it at your feet (not three feet away!) or – better yet – in your hand, so you can throw it again.  This is a good game because not only does it teach your dog to listen to the SIT, WAIT, FETCH and DROP commands, it also shows him that you are the one who initiates and ends games.  In a good game of fetch you should always end the play when the dog wants more.

Bad game: Sending your dog outside alone to entertain himself
Dogs are social animals.  Your dog craves companionship, YOUR companionship.  Sending him out into the yard (even a safe, securely fenced yard) is NOT a good way to exercise him.  Leaving him out on his own for long periods of time gives your dog an opportunity to choose his own methods of entertainment. And in many cases these will not be things you would like him to do.  This can teach your dog to bark incessantly, fence fight, dig holes, eat poop and engage in other bad habits.  It also teaches your dog that you don’t have much interest in having a relationship with him.

Good game: Take a walk with your dog or play in the yard with him
Your dog LIVES to spend time with you.  Dogs evolved to be our companions in life.  Spending quality time with your dog — time where your focus is on sharing time with him in a constructive way – is the key to building a meaningful relationship with him. 

Smart Poodles!

Irina Markova and her Performing Dogs on “Late Night with Conan O’Brien” January 8th, 2008.

Truly a pack of Very Smart Dogs!