Lulu Gets a Golden Handshake

three photos of Lulu, a young black Labrador who chose not to become an explosives detection dog.
Photos from the CIA Pupdate: A Pup Leaves the Class.

Have you ever started a job and realized, during new hire training, that you’d made a terrible mistake? Who hasn’t decided that a job just isn’t the way they want to spend the majority of their waking hours.

Well, Lulu, a year-and-a-half-old Labrador, gave up on what many dogs might consider a highly desirable career; she quit her gig as an explosives detection dog during training. Lulu was recruited from a service dog school at a young age, apparently having decided that a life of service was also not her calling. (Often, service dog puppies with exceptionally high energy or drive are released to a career like explosives detection or search and rescue, if their energy level is not suitable for work as mobility assistance dogs.)

Lulu, according to tweets from the CIA K9 training program and articles in the New York Times and Washington Post, gave up the opportunity to work 60-hour weeks with handlers from the Fairfax County (Virginia) police department. Her new life entails playing with her former handler’s children and protecting the family home from squirrels and rabbits.

Not all dogs are cut out to be working dogs. Service dog and guide dog schools that breed are doing well if more than half the carefully bred and socialized puppies actually end up working as service dogs. Some are released for health reasons, but a large number choose, as Lulu did, to just be dogs. I’ve trained lots of Lab puppies: If food and play weren’t enough of a reward to get Lulu to love the training, she really wasn’t cut out for the work. It’s to the CIA program’s credit that they let Lulu go.

“For our K9 trainers, it’s imperative that the dogs enjoy the job they’re doing,” states the “Pupdate” announcing Lulu’s retirement.

That’s a far cry (and very welcome evolution) from the “bad old days” of training, where lackluster performance was punished. Mistakes were also punished. Insufficiently speedy correct responses might also have been punished. Dogs were compelled to do the job. I am happy that more and more organizations, from service and guide dog schools to military and police dog trainers, are learning that punishment is the wrong approach.

Think about it. If compelled, the dog might do the work, but probably not put her heart into it. If your child is lost in the woods, or your city is hosting a large public event, or your city’s buses are plagued by the threat of terrorist bombings, do you want a dog who’s just doing what he has to to avoid punishment to be the search or sniffer dog on duty? Or do you want an eager dog who loves the work, buys into the goal, and puts heart and soul into the search?

It’s also cool to note that the trainer who wrote the Pupdate talked about working through a slump, figuring out what’s bothering the dog, and motivating the pups with toys and food. That sounds like they treat the dogs as individuals with preferences and feelings, not like robots who are just expected to do as they’re told. This is how it should be; dogs are individuals and should be given opportunities to make choices and express preferences.

It also raises an important point that dog trainers and owners do well to remember: The trainee, in this case, Lulu, determines what is motivating. And what is not. Most Labs love food and will do anything for a food reward. Many dogs are delighted to earn a play reward. A dog who doesn’t want to work for these rewards either needs a creative trainer to find what motivates that dog — or she needs a different goal.

Lulu made her preference clear, and I’m pleased that she got her wish. I’m betting that the handler’s children are equally delighted with her choice.

Advertisements

Bit by Bit

Her motto is “don’t fence me in”; she needs a reliable recall!

For a little over a year, I have been writing for an online “magazine” that covers online learning. Some of my articles touch on various learning theories and how these might be applied in “eLearning” or online training and performance support. Every so often, a source talks about “spaced learning” or “spaced repetition.” It’s funny to me that many of them think they’ve stumbled on a new secret weapon.

Spaced practice is simply practicing a skill or recalling information, maybe by taking a quiz or using flash cards, at short intervals over a period of time. It’s long been known to be effective both in the education of children (and adults) and in the education of dogs.

I first learned about spaced repetition in my first dog training classes. Particularly when working with very young puppies, we paired very short training sessions with spaced practice on the same set of skills. Hmmm, did we also discover microlearning, another current hot topic in eLearning? Maybe so.

Microlearning — very short lessons — certainly worked with our 4-week-old puppies. Two five-minute sessions a day (even one, if we got busy) were enough to teach one puppy I worked with about 15 cues by the time she was 10 weeks old. I’ve used the approach, with great success, with dozens of other puppies and adolescent dogs, but my training skills are becoming rusty from disuse.

I’ve been thinking about both microlearning and spaced learning for two different reasons this week.

One is a conversation I had about a friend of hers who wants to train her dog to help her with some aspects of her MS. The dog already has some foundational skills. The friend, due to her MS, has limited stamina and would only likely have enough energy to work with her dog for a few minutes a day. I found myself telling my friend that that was not an insurmountable obstacle. If she were able to spend a few minutes a day practicing a skill, the dog could learn it very well.

The second reason is my own lapse in spaced practice. When Cali was a puppy, I was very diligent about practicing her recall and working on her willingness to accept grooming, especially handling of her feet and nails. I did everything by the book, practiced almost every day, sometimes two or three short sessions. Cali was a great student. Then life got in the way and I stopped practicing. She now has a mediocre recall and hates having her feet handled. Great.

I know that I can find five minutes a day to work with Cali, and I know that, with the right treats, her recall could again become speedy and eager. That matters here in Montana, where there are many places that Cali could run off-leash.

I’m less sure about the feet, but I could probably get her to be a little less skittish and a little more cooperative. It’s certainly worth a few minutes a day to try. I have a new bag of very desirable treats; maybe I will use them for some remedial training. Wish us luck!

 

It’s Teaching, Not Training


Jana peruses several cards, then chooses "tug," asking to play tug.

Someone whom I admire greatly, though I have never met her, wrote a wonderful column a while back. It was emailed to me recently, and I’d like to bring it to readers’ attention. She describes an approach to educating dogs that she calls “Non-Training.” I’ve called this “Cognitive” dog training.

Whatever you call it, the idea is the same: Treat the dog as a partner and student, not as a robot who must obey.

It is a positive, motivation-based approach to teaching dogs that relies on their intelligence and problem-solving ability — not rote memorization of specific, inflexible responses to commands barked at them by an “alpha” human.

It starts with giving dogs choices. Teaching dogs a simple “yes” or “no” response is pretty easy. You can use your hands or ask the dog to do a different response, such as raise a paw or nose your knee. It’s up to you (and your dog) what response works. But once the dog gets the concept of a choice, the sky is the limit.

You can ask your dog whether she wants to go out, wants water, is hungry (use caution if you have a Lab or a golden!), wants the ball or the tug toy, wants a walk or a ball game, wants to rest, is hurting somewhere … Really, as in so many areas of our relationships with dogs, the limitations are imposed by our lack of imagination, not the dogs’ lack of ability or willingness.

Cognitive training goes beyond simple choices, though. It is about teaching dogs to think and solve problems, rather than waiting for us to tell them what to do. It is about shared goals, rather than humans ordering dogs around. It is a dramatic re-imagining of the human-dog partnership. It requires letting go, once and for all, of the idea that dogs “have to” obey humans, just because we have thumbs, or that they do stuff for us out of unconditional love.

A relationship with a dog is just that: A relationship. It takes work, mutual respect, a two-way avenue of communication that acknowledges what each contributes and what each needs from the other.

I won’t go so far as to say that it is or should be a completely equal relationship. Sometimes — often — the human gets to make the decisions. Much as some people loathe the comparison, it is like parenting in that way. The adult human is in charge, but that doesn’t mean there’s no room to respect and encourage the dog’s (child’s) individuality, allow for expression of preferences, and allow the dog to make choices when appropriate.

I encourage you to read the column, “Teaching Your Dog to Say Yes or No: The Art of Non-Training.”

 

A Conscious Struggle

This slideshow requires JavaScript.


Our dogs have jobs that they are expected to do, similar to assigning household chores to a child. Jana brings in the morning paper. Jana and Cali each bring me a shoe when we’re getting ready for a walk. And each dog is expected to bring me her food bowl when she’s done eating.

Jana learned this task very quickly. She does it eagerly, always happy to receive her dessert in exchange (a small cookie). Alberta is also an eager participant, usually the first to finish her meal and deliver her bowl. If I am not paying attention, she’ll push it into my leg, hard, in a not-at-all-subtle demand for her pay. Past dogs have learned the task too, though Wylie made it quite clear that he saw this task as beneath him — girls’ work — and he would immediately depart for more masculine pursuits upon finishing his meals. Jana was happy to pick up the slack (and the extra cookie).

Then there’s Cali. For months, she’d pretend to try, lifting the bowl by one edge before dropping it with a loud clang. Putting on her saddest golden retriever face, she look at me as if to say, “I tried, Mom, but I just can’t lift it. It’s a very large bowl, and I am a very small puppy.” As the very small puppy became 56 lbs. of solid muscle, this excuse held less and less water, but the sad eyes … well, let’s just say that she does a very good “sad golden” face.

Along came Deni. Not only did we switch Cali to a smaller bowl this summer, but Deni took over much of the feeding supervision. Deni wasn’t moved by the sad face. Weeks of cajoling and praise and cookies paid off. Cali started bringing her bowl, but reluctantly and only with much encouragement. Once we knew that Cali understood what was expected and that she was fully capable of doing it, the coddling was phased out.

Now, as Deni explains, Cali broods over the looming task or “stands there and glowers at her bowl.” She knows she’s supposed to pick it up; she simply does not want to. It’s similar to her reluctant acquiescence to brushing her teeth. But, more and more, Cali talks herself into bringing the bowl with no prompting from either of us. Deni reports that if Cali sees that extra-special treats are being offered for dessert, she is able to convince herself to pick up the bowl much more quickly, and she does so with far more energy and enthusiasm.

The battle of the bowl shows Cali’s increasing maturity and offers a window on her personality and intelligence. We all sometimes choose to do things that we do not want to do. Sometimes we do them for rewards (a paycheck, working extra hours to earn a vacation …) or to avoid worse consequences (dental checkups) or because we don’t want to disappoint someone we care about. Whatever the reason, we all face large and small decisions every day. Cali is no different.

Beyond WordsIn his wonderful book Beyond Words, which I have mentioned before, Dr. Carl Safina talks about sentience, cognition, and thought as the “overlapping processes of conscious minds.” He defines sentience as “the ability to feel sensations”; cognition as “the capacity to perceive and acquire knowledge and understanding”; and thought as “the process of considering something that has been perceived” (page 21).

We see all of these at work in Cali’s struggle. She understands what she needs to do and how to do it; she feels stress, anxiety or discomfort of some kind when she thinks about this expectation, and she considers how to resolve it. Sometimes the feeling of anticipation or desire for a treat wins; sometimes, probably, the desire not to disappoint Deni or me tips the balance; sometimes she needs some prodding. But she is very definitely making a choice. Whether she decides to go for the cookie or she’s concerned about our reaction or she walks away and lets Alberta or Jana pick up her bowl, she is in control; she weighs the options and makes her choice. She is not instinctively or automatically responding to a stimulus in a conditioned way, as some behaviorists would argue. Every day, every meal, Cali shows what it means to be a thinking dog.

 

Minding Their Manners

I recently read a study that compared wolves’ and dogs’ ability to solve a “puzzle” — opening a plastic box with a piece of sausage inside. The wolves did much better than the dogs, and, from the articles I read about the study, it seems that many researchers are interpreting that to mean that the dogs are dumber than the wolves, at least when it comes to problem-solving. The comments were rather unkind to dogs, and, I think, wrong.

One comment, in the New York Times, came closest to “getting it.” This person suggested that perhaps the dogs had been taught not to take human food or open food containers. Paired with the fact that most dogs’ food is handed to them by humans, while most wolves must find their own food, I’d say the dogs were set up to fail.

The study was published in a British journal by an Oregon State University researcher, Monique Udell. In her own analysis, she paid more attention to the fact that the dogs spent more time looking at the familiar human (who was present for some trials of the test and who provided encouragement in some trials) than at the box. Various commenters’ interpretations of the dogs’ looks ranged from “seeking assistance” to “slavish.”

Half the dogs tested were pets; the other half were shelter dogs, but no information was provided about how many of those had spent part or most of their lives in homes before landing at the shelter. The wolves had been socialized to humans, but even a tame wolf is still a wolf, not a domestic pet.

It makes sense that the dogs would have been taught not to take food or been punished for taking food. It also makes sense that, if a familiar human were present, they would seek help, information, or even permission before helping themselves. When I give my dogs a particularly spectacular treat (a 2-inch piece of sausage would certainly qualify), they often look at me, look at it, look at me — going back and forth a few times, seemingly questioning whether this bounty is truly meant for them. They are polite. They know the rules. They are also quite happy to indulge in exceptions to those rules, once they’re sure they won’t get reprimanded for doing so.

I would have been much more surprised if the dogs didn’t look to the human for permission or help. After all, thousands of years of living together has resulted in close partnerships and, at least on the dogs’ part, exquisite ability to read human’s communication. We humans are less successful at reading the dogs, sadly. Their survival depends on reading humans’ cues and behaving accordingly. Wolves have no such hangups (nor should they).

Some comments on the study went so far as to suggest that training dogs has made them dumber and less able to solve problems, that their social connection to humans puts them at a cognitive disadvantage. I disagree. While some training approaches do discourage dogs from thinking, modern approaches to training that use motivation and reward actually encourage problem-solving. Far from dumbing dogs down, their enhanced social sensitivity to humans enables them to thrive in our world and, in many cases, enjoy comfortable lives and strong connections with their adopted families.

And if those strong connections compel dogs to ask before eating your food, what’s wrong with that? Many people wish their roommates were as considerate.

For more about the study, see these articles:

OSU study: Have we made dogs lazier or dumber than their ancestor wolves?, the Register-Guard, Oct. 2, 2015

Why Is That Dog Looking at Me?, The New York Times, Sept. 15, 2015

The study, “When dogs look back: Inhibition of independent problem-solving behavior in domestic dogs,” by Monique Udell, was published in the British journal Biology Letters on Sept. 16, 2015.

See also my related post on the PPG Barks blog.

Canine Con Artists?

I originally wrote this post for PPG Barks, the blog of a professional positive trainers networking organization. The post was rejected; I think the reason is that I am asserting that dogs deceive each other and humans. I am  very interested in this topic, and I plan to revise the post further (or write an entirely new post) about dogs and deception. Meanwhile, I’d love some feedback from you. Please comment on the post or to me privately if you feel inclined. I am interested in what other dog people think about the question of doggy honesty and deception.

How much is a dog willing to bend the truth or improvise in order to get a reward?

That’s not a crazy question. Dogs routinely exhibit all of the cognitive behaviors needed to form an idea, plan, and execute deceptive or manipulative behavior. Consider:

Cookie, please
Cookie, please
  • Dogs deceive each other or fake each other out to get what they want. One dog will pretend to hear someone at the door and bark the warning bark — anticipating that his doggy sibling will run to the door. The conniving canine then steals the dupe’s rawhide, toy, bed, choice spot by the TV, etc.
  • Dogs who have been taught to ring a bell or bark when they need to go out tend to go through at least a short period of ringing that bell constantly … or at least testing out how often they can get Mom and Dad to “hop to it” and let them out, even when all they want to do is roll in the grass or bark at the neighbor.
  • Is there any dog who hasn’t tried to convince her owners that they have “forgotten” to feed her?
  • Many dogs will retrieve items that have not been requested in hopes of getting a reward. My dogs routinely bring me extra shoes in the morning, after they’ve been asked to bring my dog-walking shoes (and have been rewarded for doing so). This is probably optimism more than dishonesty, though. I routinely reward them for bringing me things that I have dropped, whether I was aware of dropping the item or not.

It gets even more sophisticated. For example, our German shepherd used to pretend not to know where the ball had landed when we threw it and he was busy sniffing something or chasing a squirrel. A request or two to get the ball would be completely ignored. Or, to humor the annoying humans, he’d search half-heartedly for a few seconds before doing the dog equivalent of shrugging and going back to something more interesting. “OK,” we’d say. “If we’ve lost the ball, it’s time to go home.” In under 10 seconds, he’d have found and delivered that “lost” ball.

Then there’s the golden who used the bells on the door to get Mom to open the door, knowing that her annoying puppy-sister would go charging out the door … while she stood there, smiling, as Mom closed the door with puppy outside and her inside.

So. While I will concede that not all of the above examples necessarily show deceptive behavior, some do, some might, and others at least indicate an ability to manipulate humans to obtain a desired end. I believe that dogs do lie and that they sometimes deceive each other and us. And they do it for a variety of reasons, including the possibility of getting a reward.

what the dog knowsI’ve been thinking about this since I read What the Dog Knows by Cat Warren. It’s a great book; I posted a short review here on the Thinking Dog Blog not long ago. It’s about scent-dog training, specifically, cadaver dogs. The author raises an interesting topic: False alerts. She’s brave to do this, partly because many dog people ran into what I think I am running into: Many people cling to decades-old and thoroughly debunked ideas about how limited dogs’ cognitive abilities are. But mostly she’s brave for another reason: Many handlers proclaim that their dogs are never wrong and become incensed if anyone suggests otherwise.

Some false alerts are the handler’s fault. Particularly when the handler is a beginner, and the team is at an early stage of training, the handler’s body language or other unintentional cuing might hint to the dog that “this is where” he should alert. In this case, the dog is not lying; he is trying to follow the cues he’s just learning, and thinks he’s doing what the handler wants.

Training and working in situations, like cadaver searches, where the handler is not always able to tell whether the alert is false further complicates the discussion. Some false alerts, as Warren explains, might not actually be false. She says that if they are training in a vehicle junkyard, for example, and her dog alerts on the seat of a smashed car with a shattered windshield, while that is not the target she’s searching for, she rewards the alert anyhow. The scents linger for a long time, and the dog probably did detect the scent of human decay (parts of the book do require a strong stomach!).

I’m not talking about those instances though. I wonder if — and at what stages of training — dogs intentionally, knowingly lie about detecting the target scent. There are certainly working situations where the handler might not know if the scent is present and therefore is likely to trust the dog and reward an alert. False alerts occasionally do cause problems in law enforcement.

She draws a distinction between false alerts that are outright lies and those that are more nuanced and, she says, even more insidious (though not always because of misbehavior from the dog). The dog is detecting something but is not entirely sure it’s the correct scent; or the dog has detected the scent but not found the precise location and alerts anyhow; whatever the case, in these instances, she explains, the dog isn’t consciously deciding to lie. As with human behavior, not all situations are easily explained, black or white.

Warren says she will never know whether her dog’s false alerts are inadvertent or are deliberate lies — but she does not rule out the possibility of a dog lying. She also says that her dog’s body language is so clear that she thinks she could tell if her were lying. Many humans betray their dishonesty through body language. Sometimes those “tells” are very subtle. A close study of our dogs’ body language might be our best chance at knowing when they are — and are not — trying to con us.

What do you think? Have you ever worked with a pathological doggy liar? An occasionally dishonest dog?

 

Makes Scents

Follow your nose … wherever it takes you!

In a recent conversation, a friend described her memories of her grandparents’ house — each room had, obviously, a visual memory, but also a unique scent memory. Many people have experienced being transported back to a meaningful childhood moment upon smelling a familiar scent — Grandma’s perfume, cinnamon rolls, Thanksgiving turkey roasting … But until I read Dr. Gregory Berns’s study “Scent of the Familiar,” I hadn’t really thought of memories, scents, and dogs in the same context. This study, published in the journal Behavioural Processes in 2015, shows that dogs have a pleasurable association with some scents. Berns’s team analyzed MRI scans of several dogs, comparing their brain response to sniffing the scents of familiar and unfamiliar people and dogs. The dogs were all trained, using only positive methods, to lie still in MRI machines, as described in Berns’s book, How Dogs Love Us.

What the team found was that, while the scent-processing areas of the dogs’ brains responded similarly to all the scents, only the scent of a familiar human triggered a strong reaction in the “pleasure center” of their brains. The study was carefully controlled — the scents were from people in the dogs’ households but not the dogs’ primary caregivers and not individuals who were present in the lab for the testing.

Berns compares the dogs’ ability to recognize and happily respond to the scent of a loved but distant human to the response humans show when viewing photos of loved ones who are not present.
In simple English, this study confirms that dogs, like my friend, can experience pleasant memories via scent. It makes perfect sense, since dogs’ ability to detect and identify scents is so powerful. Humans tend to rely more on visual cues to identify and remember places, people and, maybe, dogs. But dogs are far less visually oriented, and their eyesight works differently from ours — fewer colors, more emphasis on shadows and movement, for example.

It also most likely explains how dogs remember people they have not seen in a very long time. As a student at Bergin University (then called the Assistance Dog Institute), I trained a young puppy, from her birth to about 13 weeks of age. Cassie was a brilliant puppy, and we spent time together every day. Then, the semester ended, and I left for my home. I did not visit the school again until a year and a half later. I knew that Cassie was still there, but there was no way that I could pick out this now-grown golden retriever from a sea of beautiful, but very similar-looking, goldens. I’m sure I looked different too, but she had no trouble recognizing me. I quickly guessed that the enthusiastic, whole-body-wagging-with-joy girl who was leaping into my arms was Cassie. The nose knows!