Horses, Dogs, Learning, Serendipity, and Motivation


The material for this month’s commentary – horses, dogs, learning, serendipity, and motivation—began coming together in my mind years before it found its way into words. But then, sometimes that’s the way learning and communicating what we learned  work.

It began with a short book I received in late December, 2014 entitled the Nature of Language by Gertrude Hendrix. (Research and Review Series, Eastern Illinois University at Charleston. No.1, 08-1988.) In the letter accompanying it, my client admitted that parts of the book were heavy reading. However, he thought I might appreciate the animal anecdotes the author, an avid horsewoman, used to support her learning hypotheses.

Work and life intervened, but the slender volume remained at the top of the pile of books and journals I intended to read when time allowed.  Or so I told myself. In fact, the title of Part I—”Unverbalized Awareness as an Agency for Transfer of Learning”—and skimming that revealed an equation with a symbol WordPress doesn’t even recognize defeated me.  Both  clearly communicated I wasn’t ready to read the book yet.

In May, 2019, I overcame my aversion to reading the book out-of-order and read the equine anecdotes as my client recommended. In these, Hendrix describes the cross-species magnitude of the learning hypotheses those horses led her to consider. Encouraged by what I’d learned, I then read the rest of the book. I loved the book so much I reread it in order from beginning to end.  And no doubt I’ll  read it again.

In Part I, Hendrix describes her serendipitous, simultaneous pursuit of 3 quite different endeavors  in 1937:

  1. Early each morning she schooled a difficult horse.
  2. She was teaching a 9th grade algebra class in a teachers college laboratory school during which she attempted to demonstrate an inductive teaching approach to observers. (Inductive reasoning involves making generalizations based on behavior observed in specific cases. Compare to  this to deductive reasoning which uses given information, premises or accepted general rules to reach a proven conclusion.)
  3. She also supervised a student seminar examining the literature on the transfer of training.

During one training-related seminar, a student wondered if  subconscious generalization of learning could occur. No literature on the subject existed then, but Hendrix couldn’t get that possibility out of her mind. How might it relate to the horse she’d spent 3 unproductive years trying to train to canter instead of bolting in an indoor ring? About that same time, she had decided to try a different training approach on the horse: She trained him to canter in a small ring where he had no room to work up enough speed in a canter to bolt. When he learned the desired behavior she rewarded him. After a while he learned to canter normally in a large ring, too.

By May 1937, the horse performed so reliably in the ring that Hendrix decided to test his training on a 6-mile ride on carefully chosen country roads. The course included four level stretches, each with a steep hill at the end. She reasoned that, if  the horse did bolt on a level stretch, she could use the steep hill at the end of it to regain control of him.

Everything went perfectly during the first two stretches during which the horse walked, trotted, then cantered as trained. But as they came around a corner and the horse saw the third level stretch with its steep hill at the end, he immediately put himself into the position to canter. He was so excited about what lay ahead that Hendrix feared the high-energy young animal would lose his head, break out of the canter and into a gallop. If he did, she’d lose everything she’d gained in months of conditioned training.

Once again, serendipity stepped in. Hendrix chose not to dwell on the failure of those months of tedious horse-training lessons. Nor did she bemoan the effort required to hold an increasingly agitated horse to a walk for the remaining 3 miles lest he hurt himself as well as her. Instead, she thought about what her horse had done in terms of learning theory. He had generalized the pattern he’d learned in the first two examples in the new environment and wanted to apply it to the third. This internal and thus highly motivated generalizing had cut through months of unmotivated training based on conditioned learning hypotheses fueled by apple slices doled out at the appropriate intervals.

I’ve seen a similar phenomenon in intelligent trained dogs with behavioral problems many times. Like Hendrix’s horse, all these dogs were intelligent. While the owner was thinking in terms of command-response-reward or punishment, the dogs were processing that same information inductively. And then they were making generalizations from it, also sometimes in remarkably short periods of time. Like her horse, these dogs also would self-initiate training sequences taught by their owners. But unlike Hendrix’s horse who remained under her control the whole time, the majority of these dogs were off-lead when they did this. And none of them had a reliable response to the come command to stop them from completing the sequence. This created a serious dilemma for their owners: Should they reward the dog or not?

Hendrix avoided that learning vs. teaching/training dilemma. However, the realization that her horse could generalize what he learned caused her to changed her teaching approach. Her new one took into account the animal mind’s ability to process information in two ways not addressed in conditioned learning hypotheses:

  1. It could handle sequences of consistent “if” and “then” events—i.e. if my owner gives this (verbal or nonverbal) command, then I do that.  And then that. Although training assumes the  animal translates the “if” translates purely as, “If my owner/trainer gives the command  x I do y”, the animal’s “if” includes a lot of other factors.
  2. The animal mind doesn’t just learn how to do what the trainer wants the animal to do. The animal mind also can generalize new knowledge and apply it to other situations, often in a short amount of time.

Hendrix also noted that many of the things she did following this awareness would have been indistinguishable compared to those she did based on conditioning hypotheses. What changed was the motivation that fueled the trainer and the animal. And that changed their bond with each other. She wrote:

Certainly, the trainer’s regarding each new thing learned as awareness of a generalization elevated the equine learner to a position of higher respect in the trainer’s regard. There was a kind of pervasive dignity between trainer and horse that had not been there before, a kind of empathy which generated a great respect in the trainer for the horse and, very soon, a seemingly greater respect of horse for trainer.

Mutual respect, pervasive dignity between animal and trainer, empathy. Plus knowledge-based serendipity and awareness of motivation. Sadly, all concepts that seldom may arise in discussions of how to train companion animals.

Leave a Reply