Technical Writing as Dog Training

I have two Shetland Sheepdogs (Shelties): Miso, a five-and-a-half-year- old sable male that looks like a miniature Lassie, and Mochi, a black and white female whoʼs a year younger. Both are graduates of the Saratoga School for Dogs.

The Saratoga School for Dogs teaches dog obedience. Thatʼs not accurate. You canʼt drop off an untrained dog, then pick it up obedient ten weeks later. What the Saratoga School for Dogs actually does, is teach owners how to train their dogs. It trains the trainers.

When I enrolled Mochi in the school, Iʼd been through the training the year before with Miso. This gave me an opportunity to think more about the training, and less about whether Mochi would graduate magna cum canine. I knew she would. As I listened to the instructor explain his training philosophy, it struck me how much dog training was like technical writing, or any kind of communication where the goal is to enable a learner to complete a specific task.

Proof of concept

Every day, I walk Miso and Mochi to Starbucks. If I say “Heel,” they line up on my left, Mochi next to me with Miso outside her, their noses aligned with my knee. They adjust their walking speed to mine. If I run, they run alongside in formation. If I suddenly stop, they stop with me. If I turn to either side, or in a circle, they continue keeping pace. When I stop at a curb, they stop, then after a moment, sit, awaiting further commands. If while weʼre walking I say “Stand. Stay” they stop in place, even if I continue walking. From any distance that they can see or hear me, if I command them to “Come,” they run to me, stopping at my feet, then sit, awaiting further commands. If I say “OK,” they are free to stop heeling, and proceed at their own pace, exploring as they will. They donʼt stray far.

At Starbucks, I show them a place outside, then command them to “Sit. Down. Stay.” They remain lying there while Iʼm inside. When I return, they remain in place while I give first one, then the other, some cappuccino foam.

People tell me they wish their children were as well-behaved.

Smarter than you think

The first thing to know is that dogs (and humans) are smarter than you think.

Dogs have extraordinary learning abilities. They can be trained to assist the handicapped, guide the blind, help police and the military, protect our homes and children. Most people know at least one example of amazing dog behavior. But even mundane examples suggest thereʼs more to a dogʼs mind than might be imagined.

When I was training Mochi to come on command, I repeatedly told her to “Sit. Stay,” then walked a few feet away, and said, “Come.” When she completed the behavior correctly, Iʼd repeat the commands. After dozens of iterations, I became confused, and instead of saying “Come,” I said, “Sit.” She looked at me, her hind-quarters tensing, then relaxed, and remained sitting.

Itʼs clear Mochi was not blindly repeating a sequence of actions. She understood the difference between “come” and “sit.” She was parsing each command separately. She resolved a discrepancy between expected and actual commands.

Frequently, after giving Miso and Mochi some cappuccino foam, I leave them and walk away. They dutifully remain lying down. At any distance where they can see or hear me, I can say one of their names — or point to one of them — then give a command, such as “Come.” The designated dog will come to me while the other remains lying down. I can do this in any order, specifying either dog, and any known command.

Clearly, Miso and Mochi have at least a rudimentary notion of sentences composed of subjects (their names) and verbs (commands). If I do not prefix a command with one of their names, or prefix it with “Miso, Mochi,” they both respond simultaneously.

(Miso and Mochi also know the names of our three cats: Sushi, Tofu, and Mocha. If I say one of the catʼs name with emphasis — “Sushi!” — both dogs run to confront the offending cat. The catʼs name said without emphasis elicits no response.)

Mochi barks at dogs on TV. She does not bark at other quadrupeds, such as cats, horses, or even sheep. Only dogs. She does this even when the sound is muted, meaning sheʼs responding to the image of another dog, and not any sound it makes.

Whatʼs more interesting, Mochi barks at cartoon dogs. This represents a level of abstraction; a generalized notion of another dog in Mochiʼs mind. Again, she does not respond to other cartoon animals.

Dogs can learn obedience and at least 150 behaviors. They are smarter than you think. If learning doesnʼt occur, itʼs not because the dog isnʼt smart enough, but because the trainer failed to train it.

So it is with human learners. When I was earning my Secondary Teaching Credential, our instructors had a slogan: “If the student hasnʼt learned, you havenʼt taught.” It made no difference what techniques were used, or under what condition teaching was attempted. If the student hadnʼt learned, we hadnʼt taught.

The geneticist, Arthur Soller, once explained to me he prepared for lectures by assuming his audience was “infinitely bright, but ignorant.” His job then, was the simple eradication of ignorance. The audience was always smart enough.

There are no dumb users.

Different Senses

Dogs are well-known for their extraordinary senses of smell and hearing.

I have watched Miso, his eyes closed, nose probing a breeze, luxuriating in an olfactory world invisible to me. The mentation required to analyze and correlate those bits of smell data with things and events in other places and times is amazing. It results in a mental representation of the world beyond that available to immediate perception.

Miso and Mochi can hear my normal speaking voice, even when it approaches a whisper, from at least twenty-five feet away. Even when they are sleeping at the other end of the house, they come running into my office — hoping we will finally go outside to play — when they hear the subtle change in pitch of my laptopʼs cooling fans as I set the machine to “sleep.”

Why then, do people yell at dogs to get them to obey? Dogs have excellent hearing, more acute than ours. If they are not doing what we asked, itʼs because, well ... We havenʼt trained them. They hear perfectly well.

By contrast, humans have better visual acuity than dogs. Vision is our specialty. We process the photons of our world as exquisitely as dogs process the molecular interactions of theirs.

Why then, do we visually yell at human learners? Are all capital letters, difficult- to-read underlining and italicization, bolding, “ransom-note” typography, and other non-semantic formatting easier to comprehend than simple unformatted text?

When humans train humans, we donʼt contend with wide inter-species differences. Nevertheless, there are both cultural and physical differences in how humans perceive the world. We often do not appreciate the learning problems of others who are deaf or blind, near- or far-sighted, color-blind, or even illiterate. We are all trainable — infinitely bright, as Dr. Soller would say — even if we do have to take our differences into account.

Treats

Treats are not used for training at the Saratoga School for Dogs. The reward for correct behavior is for trainers to “Praise the heck out of your dogs;” an enthusiastic and sincere, heart-felt “Good dog!” with lots of petting.

Thereʼs good reason for this. If for example, you and your dog are in the front yard and it runs toward a squirrel across the street — just as a speeding car approaches — you donʼt have time to tempt it with a treat. On the one occasion when this precise scenario occurred, I called “Stay!” and Miso stopped dead in his tracks, before he left the curb. I praised the heck out of him.

You can always praise your dog. You donʼt always have a treat in your pocket.

Thereʼs another reason to use praise as reward: It works. Itʼs what motivates dogs — and people. To varying degrees, all dogs want to please their owners; do a good job; get that “Good dog!” with lots of petting. Similarly, no human shows up for work with the intention to fail, or do a bad job. No one wants to be criticized for poor performance. We all want to succeed. It feels good when we do.

The best reward you can give a learner is to enable them to do their job well. Provide them with useful information. You donʼt have to finagle users into learning. Itʼs exactly what they want to do.

Distractions

Miso and Mochi obey commands under virtually any condition. When we walk to Starbucks, we walk through urban areas. There are stores with people coming and going, children on mechanical rides, cats wandering around a pet store, dogs, birds, and other animals, cars, trucks, and occasional loud music. (Mochiʼs easily startled and frequently frightened by both loud and low-frequency noises. In spite of her fears however, she remembers her training.)

Similarly, humans — especially online help users — are not sitting quietly in a training center, focused on a single task. They are immersed in their own chaos of instant messages, emails and scheduled tasks dinging them, telephones ringing, loud talkers, disengaged coworkers wandering into their cubes, iPods plugged into their ears, video-on-demand, and other multi-tasking diversions. Add to that help systems with animated GIFs, Flash buttons, pop-up windows, hide- and-show content, and the latest technological tour-de-force.

Oh: Theyʼre presumedly also trying to complete some work-related task.

And like Mochi, users may be fearful: of failure, loss of status, and the like. Online help users are especially likely to be anxious because online help is generally considered a last-ditch effort, consulted only when all else fails.

Technical communication has to take into account learnersʼ real-world environment. Remember: Your help system — or manual, quick start guide, demo, or what have you — isnʼt the only thing competing for usersʼ attention.

Focus

One of the first exercises at The Saratoga School for Dogs is to let the dog sit — they do so naturally — then wait until it looks up at you. When it does, praise the heck out of it.

Dogs always look at their ownerʼs faces for information, especially if they need help. Dogs differ from most animals in this respect, and in following their ownerʼs gaze. They are interested in what interests us. (Iʼm always astonished to realize Miso and Mochi are observing my behavior, as much as I am theirs.)

When writers embellish their products with visual and aural distractions, they are trying to focus learnersʼ attention, but in the style of those obnoxious blinking, flashing, constantly moving web advertisements we all soon learn to ignore. What learners really want is help; less flash, and more substance.

Technical writing should be recognized by learners as useful. Something that helps them get their jobs done. Something that helps them succeed.

Think of the first time someone uses your online help system for example, as your dog looking up at you. Give them something they need, not what pleases you.

Anticipation

When the dogs and I go for our daily walk, we can turn left or right from our house. I used to alternate the direction each time, until I noticed Mochi charging ahead, turning the opposite way weʼd gone the day before. I then changed the rules, choosing the direction opposite the one Mochi chose. Now she waits until I make a choice before following me and Miso.

No overt training was involved. Mochi and I simply observed each otherʼs behavior, then anticipated what the other would do. I only “won” because Iʼm the bigger dog.

Humans can be even better observers of one anotherʼs behavior. A good part of social interaction depends on predicting how another human will behave. Itʼs what distinguishes the socially adept from their opposite.

Anticipation however, can be an impediment to learning. Teachers, especially if theyʼve taught the same class for a long time, often answer anticipated questions rather than ones students actually ask. In rapid conversation, we anticipate what someone will say, interrupting them with our response before completely hearing what they said. Learners can have preconceived notions about a subject that effectively prevents them from seeing it in a new way. Trainers need to be aware of this, both in learners and in themselves.

I once had a sign over my laboratory bench that read: Maybe itʼs good Iʼm going so slowly, because I might be going in the wrong direction.

Hear that, Mochi?

Whatʼs the trick?

The critical part of learning, for dog or human, is separating what is pertinent, from what is extraneous. For example, when I was training Mochi to come forward to have her paws cleaned, I unconsciously motioned her forward with my hand. Now she stares at me waiting for the hand motion before obeying the command. (I decided to accept this behavior, rather than correct it. Perhaps itʼs Mochi who trained me!)

If you always yell when training your dog, it will learn to obey only when commands are yelled. If you start a command by settling yourself with a softly spoken “OK, now,” that phrase becomes part of the command. How does the dog know that raising your hand to your cheek, or jiggling your car keys, is not part of the command?

Human beings fair no better. When asked why they complete particular tasks in inefficient ways, users often explain thatʼs how they were first taught. Itʼs well- known that young children persist in doing tasks the first way that succeeded, even when the method for success changes.

Think about this, especially when making demos; when cursors are flying across the screen, menus and dialogues appearing and disappearing. Think about this when making screenshots of complicated dialogues. Assume every movement, check box, option button, drop-down selection, and so forth is being learned as part of the trick. If itʼs not, donʼt put it in there.

Minimal Vocabulary

One way we differ from other creatures is we have a lot of working memory: RAM. We can hold a lot of variables in our head at the same time. People who do not have this ability, either through injury or birth, find adding large numbers together in their heads difficult, although they can do the math on a piece of paper, using it as a type of external memory.

While dogs have much better brains for some things than do we, they have less verbal-RAM. They canʼt store as many word variables simultaneously. Dog training is best suited for a minimalist vocabulary.

When I tell Miso and Mochi to “Come,” I always say it exactly the same way. I donʼt sometimes say “Comʼon” or “Comeʼre” or any other variation. I want to conserve their verbal RAM.

People have the same problem, but obviously to a lesser degree. I was once confounded by simultaneous problems reported for the Moyno pump, Progressive Cavity pump, Eccentric Worm pump, Tank Farm pump, Molasses pump, Number One pump, and Fermentor Feed pump, only to discover all eight were the same pump. High-tech is especially egregious in this way, often preferring to invent a new term rather than use an existing one to describe the same thing. Sometimes the Training Department uses a completely different vocabulary than the manuals. Occasionally, programmers use yet another vocabulary for error messages, and other on-screen text.

I know a minimalist vocabulary works for training dogs. Iʼm sure it also works with humans.

Theory of Mind

In a classical experiment, a child behind a one-way glass observes an experimenter placing a toy inside a box. The experimenter leaves, and another child enters the room to retrieve the toy. If the observing child is young enough, he will be puzzled if the other child does not know where the toy has been hidden. He knows where it is. Why doesnʼt the other child have his same knowledge?

This behavior usually disappears about the time children begin school, and they are no longer puzzled that other people donʼt know whatʼs going on inside their heads. They begin to have a notion that other people have other thoughts. They are said to have developed a “Theory of Mind.”

High-tech seems particularly afflicted with a defect of Theory of Mind. We are so close to our subject we find it inconceivable anyone would not know what we intended our software and hardware do. When we say something is a “user problem,” are we really saying users donʼt know whatʼs in our minds?

Conclusion

At least 30,000 generations of dogs have lived with humans. Weʼve gotten to know each other fairly well. We might even imagine we know something about what goes on in one anotherʼs minds. We certainly have had lots of opportunities to study each otherʼs behavior.

Training a dog provides a fresh look at what training really means. Itʼs a way of removing preconceptions; thinking anew about the problem of one mind communicating with another. It reminds us that at some level, all minds are different, yet similar.

One might ask, If you canʼt train a dog, what makes you think you can train a human?