Re: Associative learning versus imitation - JoM Article

Bruce Howlett (
Tue, 20 Oct 1998 01:22:00 +0000

Date: Tue, 20 Oct 1998 01:22:00 +0000
From: Bruce Howlett <>
Subject: Re: Associative learning versus imitation - JoM Article

 That was fast - nearly an "on line" chat happening here!


On Mon, 19 Oct 1998 22:07:26 +0000 Bruce Howlett
<> wrote:

> Dogs are easier to train because they
> recognise us as fellow predators (binocular vision and smell)

I'm not sure I follow this.  Do you mean that dogs are capable of
classifying other species into predators and non-predators on the
basis of recognition of binocular vision and our smell?  I would
imagine that humans probably smell much like other primates, many of
whom eg. gorillas and orang utans are not carnivores/predators.

The short answer is yes, I do believe this but as I mentioned, this was a non-scientific observation.  Any prey animal will demonstrate immediate signs of discomfort and anxiety if "eyeballed" by a predator.  A technique used to catch shy horses is to avoid looking at any part of the horses anatomy, using the position of ones body to manoeuvre the animal into a convenient position.  Dogs eyeball each other and humans, and I believe they are aware on some level of a similarity in our visual perception.  As for smell, we all smell of what we eat.  Meat eaters smell of meat, grass eaters smell of grass.  Simple.
> The key to this is understanding what
> "buttons" to push to get a horse to behave the way we want rather than
> the way they want.  This is basically achieved by making the desirable
> outcomes easy and the undesirable outcomes difficult

Yes, but what you're describing is individual learning by the horse,
co-ordinated by a skilled reinforcement from the trainer.

   > Ask any experienced rider and they will
   > tell you about the extraordinary and euphoric feeling of being in total
   > sync with a horse performing a difficult manoeuvre, whether cutting out
   > a steer or a dressage test or a cross country event.  This is high level
   > communication of a sort more subtle, continuous and cooperative than
   > mere language.

   Yes, but is there any imitation?  I don't see how a horse could even
   begin to imitate a human, as the anatomy is so different.  For
   instance, how could a horse, even a superintelligent one, imitate me
   fumbling for my keys?

But what we are asking the horse to learn is simply an extension of its own natural behaviour, modified to suit the activity specified by the trainer.  A horse can only do one useful thing for a human, that is to move (under control) in a specified direction.  Good trainers (not rough riders) get the horse to perform every desirable movement  on the ground before mounting, that means all that remains is to teach the horse to compensate for the shift in the center of gravity and to trust the riders judgment regarding distances.  Horses having bi-lateral vision cannot judge distance except by experience.  Most actual riding control is achieved through the horse imitating the center of gravity shift initiated by the rider.  The so called "aids" are used to exaggerate the gravity shift so that the horse understands that it is a deliberate action.  The classic position of an inexperienced rider on a bolting horse is collapsed forward from the waist and hanging on by the heels, which is basic horse speak for "go forward as quickly as possible"!  Most of my horses will work entirely of "seat" commands, in other words, move in the direction of a weight shift at a speed specified by the magnitude of the shift.  To me, this appears to be the horse reproducing or imitating what I am doing.
> The similarity between horses, dogs and humans is that they are all
> "social" animals with distinct hierarchies that recognise dominance and
> subservience.  Training is a much higher level activity than imitation,
> it is actually communication.

Nope.  Got to disagree there.  Many animals communicate, such as
bees, vervets etc, who show no evidence of any ability to imitate.
Imitation is no lower than communication - or indeed higher - they
are just two different things.

I sometimes think the Australian Blue Cattle Dog is the toughest animal alive. They are physically tough as well as mentally tough and almost impossible to train if they do not accept the trainer as being higher on the "pecking order" than they are.  I have retrieved many Blue dogs from the local dog pound which is where they usually end up if they win the dominant role in a family situation and become uncontrollable.  Such "alpha" dogs are highly sought after by  security companies as guard dogs and work well for experienced handlers who know all the tricks.

As regards their ability to imitate, I would suggest that a male dog lifting its leg to pee is an imitated behaviour.  I have one male dog who was raised with only bitches and I have never seen him lift his leg, he only squats like the females.  When training a dog to work stock, the handler literally works the stock on foot the way he wants the dog to work them and encourages the natural tendency for the dog to "herd" the stock the way it would naturally if living wild.  Of course the dog must be taught to do this on command.  All the good working dog trainers will tell you you must work as a team with your dog.  Handlers who expect the dog to do all the work while they hang back usually get poor results, although I have know one dog who could muster a 500 acre paddock while the owner waited at the gate.

> I simultaneously laugh and groan when I hear you suggest that the
> intellectually challenged bird is considered the only bona fide imitator
> in the animal kingdom.

But nevertheless, that is the case.  Quite how birdbrains manage to
imitate, is, I agree a bit of a puzzle (or a lot of a puzzle) but
apart from the largely anecdotal evidence in primates, birdsong
imitation is the only real model system we have for imitation outside
of humans.

I suppose this depends on what you are prepared to accept as "proof" of imitative behaviour.  I once had a Border Collie bitch who would/could not swim.  I nearly drowned her several times trying to teach her with no success.  Then I got an old Blue dog from the local pound (in fact he was the first, this was 20+ years ago) who loved to swim and would at every opportunity.  In no time at all, a matter of a few days, the Border Collie bitch was happily swimming with the Blue dog.

It works for me, perhaps not "scientific" enough for some!



Bruce Howlett.  B.A.L.,  J.P.
Researching:  Management of Change in Organizations:  The Culture Concept
at the University of New England
Armidale NSW 2350

  =============================================================== This was distributed via the memetics list associated with the Journal of Memetics - Evolutionary Models of Information Transmission For information about the journal and the list (e.g. unsubscribing) see: