Where’s the On-Off Switch on This Dog?

We’ve now had Tajji for almost three weeks and have a pretty solid sense of the challenges we’ll be facing while re-socializing her. As I noted in my first blog post about her, she’s “hyper-vigilant:” aware of and alert to anything new around her, and convinced that it is her responsibility just as it was with her blind person. This attitude is certainly a necessity for a seeing-eye dog in her service harness, but it’s a serious liability in a companion animal at the end of a leash. It manifests itself most distressingly in Tajji’s rapidly escalating response to other dogs that culminates in an explosion of frustration, fear, and aggression that’s basically uncontrollable.

Tajji gets her basic clicker training: Look, click-YES!, treat. Rinse and repeat.

Tajji gets her basic clicker training: Look, click-YES!, treat. Rinse and repeat.

This kind of behavior, whether in dogs, two-year old toddlers (the mental level of a mature dog), or adults with anger-management issues, is often called “going ballistic.” This is actually a quite useful metaphor in dog training. Just like a projectile in its ballistic phase—freefall—a dog in a frenzy is following a pretty much fixed trajectory. There’s not much you can do in terms of remediation until you remove the stimulus that set your dog off in the first place. On a walk, this generally means “get the hell out of Dodge.”

What absolutely, one-hundred-percent-guaranteed won’t work at this point is force. “But,” you may object, “I’ve seen it work on TV.” Actually, you haven’t, although, given the overwhelming dominance of flawed pop-psychology explanations of dog behavior and cognition we’re exposed to, it’s not surprising you thought you did. After all, force training the sort of problem Tajji has appears to be a very clear, three-step narrative:

  1. Dog walking with trainer encounters another dog and exhibits undesired behavior (e.g., lunges, barks, snarls)
  2. Trainer applies escalating force to stop behavior (which doesn’t match any definition of “whispering” that I know of)
  3. Dog ceases to exhibit undesired behavior towards other dog.

That’s correct, as far as it goes. But what was the dog actually thinking? What was it “saying” to the trainer during the process? Most important: what did the dog actually learn?

Once you answer those questions, the typical YouTube force-training video looks completely different. To understand why, let’s look at dog brains, how they work, and how to work with them instead of against them. There’s a lot of information to cover, so this will be a two-parter; I’ll leave you with some reading suggestions in the meantime, and you’ll have Deborah’s post next week to look forward to! The Never-Ending Terrible Twos

Dogs, like toddlers, are always struggling to engage their cerebral cortex. That’s what makes the human twos so terrible: the beginnings of a battle royale between the emotional limbic system and the executive cerebral cortex that controls and directs it that takes about 25 years to complete in human beings.

Although a mature dog is stuck at two years old, mentally and emotionally, that doesn’t have to be terrible. The good news is that dogs are much faster learners than toddlers. The bad news is that they are equally complex beings, with equally rich, if not richer, emotional and communicative lives. They are not programmable meat puppets, nor poor reflections of their human owners. To the contrary, they reflect perfectly the true behavior of their humans towards them—one cannot lie to a dog, nor will it ever lie to you—which is why so many dogs are so badly behaved.

The stark reality of dog training is that, like a toddler in a melt-down, a dog can’t learn when it’s operating from its “limbic” brain, that complex of structures responsible for “fight or flight” and related emotions. Our job is to keep dogs operating from their cerebral cortex, where they can learn, and to teach them how to get back there when they fall into limbic. Fortunately, dogs really don’t like full limbic arousal; it’s an extremely vulnerable state.

Again, force can’t help the dog here, as anyone knows who’s been foolish enough to try it with a toddler in full meltdown. Revisit for a moment the kind of force-training video described above, then shoot the “toddler” at the other end of the leash full of meth, give it the strength of a 200-pound human athlete, and equip it with a jaw full of teeth that can break your arm. Makes juggling chain saws look fairly safe by comparison now, doesn’t it? Just call it the magic of a domestication process that has made dogs far better at living with human beings than human beings are at living with dogs.

Working for Someone Who Rarely Listens

Being tied to beings who rarely listen to you is the fate of most dogs. Which is sad, because they don’t have any “internal censor,” and they’re constantly letting it all hang out. Unfortunately, little of this involves sound, so important to us monkeys. For a dog, the semiosis of sound is almost entirely focused on emotional states, so almost all of what they hear from us is simply barking to them, a fortiori when we’re angry or distressed. Instead, the dog calls upon a large repertoire of postures, positions, movements, and even more subtle visual signals. As far as a dog is concerned, she is constantly telling her human exactly what she’s thinking and feeling. (Deborah addressed some of these in her blog post last week, with some recommended reading on dog language.)

Tajji, for instance, signals her dog arousal with a set of escalating behaviors. Her ears point forward, and her movement becomes subtly stiffer. Her breathing accelerates and becomes harsher. She starts forging ahead on the leash, and pays less attention to me. Food becomes of less interest. All this happens before she even sees the other dog, as when it’s behind a solid fence. At this point it’s still pretty easy to bring her back down, away from limbic, with some focused activity such as “puppy pushups:” sit, down, sit, with a happy “yes!” followed by a treat for each movement.

However, once she sees the other dog, even from blocks away, she rapidly loses control. Quiet “whoofing” rapidly escalates to loud barking, pulling on the leash becomes lunging, and her appetite completely shuts down. As with all dogs, we have about two seconds between the melt-down stimulus, the sight of another dog, and full-on limbic arousal: two seconds before learning shuts down. At this point in her re-socialization, the only thing to to do is turn around and get her away from the other dog.

If she goes ballistic, we’ve failed. Remember, it’s never the dog’s fault; she told us from the git-go how she was feeling and what she was thinking. It’s our job to learn her language, and recognize that she has no need to learn ours, since she can already read our emotions flawlessly.

Now think about that video again. Look at the curve of the dog’s backbone, the set of its ears, the wide range of escalating placatory behaviors (such as a nose lick or sniffing the ground) that likely begin the moment the force-trainer takes the leash. The dog was telling him all along what she was feeling and thinking with increasing urgency, and offering different behaviors to try to placate the monkey, and the trainer didn’t listen, or more likely didn’t care. Either way, not a very reassuring situation for an animal selectively-bred for millenia to care very much about what humans are feeling and thinking.

What’s actually happening is that the force trainer is “flooding” the dog: presenting it with so much incoming information that it can’t process it. What looks like eventual submission (walking by the other dog without lunging, for instance) is actually surrender, shut-down, learned helplessness. What the dog has actually learned is that other dogs mean unpleasant stimuli no matter what it does. If you have a really smart dog, she will eventually learn that “the monkey at the other end of the leash goes ballistic when he sees another dog,” and it will try to work out a way to prevent that, which likely won’t coincide with your ideas of ideal dog behavior.

Next time: I Think, Therefore I’m Calm, Karen Overall’s Relaxation Protocol

Suggested reading


About Dave Trowbridge

Dave Trowbridge has been writing high-tech marketing copy for almost thirty years. This has made him an expert in what he calls “pulling stuff out of the cave of the flying monkeys,” so science fiction comes naturally. He abandoned corporate life in 2007 — actually, it abandoned him — but not before attaining the rank of Dark Lord of Documentation, a title which still appears on his business card and serves to identify clients he’d rather not work with (the ones who don’t laugh). He much prefers the godlike powers of a science fiction author (hah!) to troglodyte status in dark corporate mills, and the universe is slowly coming around to his point of view. Dave is currently laboring over the second edition of the space-opera series Exordium with his co-author Sherwood Smith, and looking forward to writing more stories in that universe. He lives in the Santa Cruz Mountains with his writer wife and fellow BVC member, Deborah J. Ross, and a tri-lingual German Shepherd Dog responsible for three cats. When not writing, Dave may be found wrangling vegetables—both domesticated and feral — in the garden.


Where’s the On-Off Switch on This Dog? — 12 Comments

  1. Great blog. My labradoodle (or perhaps standard poodle) was a highly energetic puppy. We still call him the ever-ready labradoodle, and highly reactive to stimulus, although not so much to other dogs. He also has a dominant, non-biddable personality–we call him amiable dominant. He expects to run the show but won’t bite you if you insist on something else. 🙂 I learned early on that if I was trying to correct a behavior, harsh methods (by which I mean a louder voice or leash jerk–nothing harsher) acted like a cattle prod–energizing him even more. The more we could calm our body language and voice, the better he was. And once I discovered clicker training, where he thought he WAS running the show, he really learned things quickly.

    I can’t find the link, but there’s a great video of desensitizing a highly reactive dog so that he learned to ignore skateboarders and bikers. It took some weeks, but at the end, the dog simply ignored the skateboarders and focused on the rewards.

  2. How did that work for Tajji as a guide dog then? Isn’t that sort of seriously escalating response a really bad scene for a blind person, or does she not do it when she’s in her work clothes?

  3. Alicia, when she was working, she was totally focused on her person and her job. She didn’t react at all to other dogs. The problem is that she never learned ordinary-dog manners. She’s so “soft,” sensitive, that she needs equally sensitive handling. Looking at how splendidly she’s adapted to the cats, I’m hopeful that once she gets the right structure, she’ll learn to calm behaviors instead of OMGIAMSOEXCITEDICAN’TSTANDIT behaviors.

  4. I used the Overall Protocol extensively when rehabbing Connery from a series of giant breed attacks. Most excellent stuff.

    Also Leslie McDevitt’s Control Unleashed games, which provided us with structures so he could learn to manage his own fears and behavior (although he still depends heavily on me). To this day those around me at trials have no idea how much moment-to-moment support he gets–it’s all subtle and ongoing–or how much I’m protecting him from their dogs’ nonverbal pushiness. But within those parameters, he’s been able to relax and have his life back.

    Looking forward to the next Tajji report!

    • That’s good to hear, Doranna. She went about halfway through the first day in a few minutes, according to Deborah. I haven’t worked her on it yet, as I’ve been the walkies person whilst Deborah’s knee is ouchy. There are plenty of desensitization opportunities on those walks, and if things get to be too much, we just turn around (freudische!) and go back.

      Incidentally, I noticed something today about Tajji’s existing training that speaks to the excellence of Fidelco’s approach: she does not assume that the click-yes (or just yes) for a specific behavior is a general release. That’s “OK” or now “Frei.” She’s so solid on this that she’s retraining me!

      • I’ve found with horses that keeping them moving (and with a horse you have the option of asking them to move over with their front or back legs) is an important tool in breaking the locked concentration – you could try to walk her in small circles and figures of eight or get her to weave in and out of obstacles and see whether that helps.

        Another tool – and this comes from Feldenkrais training – is to try and give them access to a _different_ behaviour. When they are locked into a reaction cycle of ‘stress-full concentration-bad behaviour’ (whether that’s running away or attacking) you have little opportunity to influence them, when they have several behaviours to choose from you get a moment where they have to actually THINK – and in that moment a human can interfere and provide guidance and distract them.

        I’ve never had a dog to test this theory with, but I wonder whether asking her to not just sit but do something – eg lifting a paw – might help.

        • That’s the point of the relaxation protocol: to give the dog a set of behaviors that you and it can use to help it calm down. They really don’t want to be that excited; they just can’t help it.

  5. That’s great. I’ve been working with a dog who understands praise and/or confirmation as a release, and I get tangled in it all the time. I’m not honestly sure (at least with this particular dog) that the release response can ever truly be eliminated. Under certain circumstances I can get away with clicking her–(I use clicking as a bridging reinforcement as well as click=cookie)–but never with verbal.

    • That sounds right. I had that problem with my last GSD due to sloppy training on my part. Fortunately, it’s not much of a problem with Tajji. Her early training was apparently splendid.