Positive reinforcement: Practical applications

Pale Rider said:
I think that food can be very powerful, but, using it in an untargeted or semi random way is a mistake. It can lead to mouthiness, invasive behaviour and a lack of concentration for the horse. I suppose the attempts to buy affection from the horse with food are the most ineffectual uses of this training tool.
Click to expand...
At best, it is a wasted opportunity in my opinion; at worst, it is letting the horse down - even though it is almost always done with good intentions.

Withness, is an interesting concept and is perhaps the acid test of a relationship. The horses willingness to remain with the handler when the head collar comes off in the the field demonstrates how successful the training regime has been, and exactly how strong the relationship is.

If the horse will stay with you in 20 acres, without being bribed with food, and if you can draw the horse from 100 or 200 meters, without a food bribe, shows that you're getting something right. These are the yard sticks I like to use to measure the effectiveness or not of the training in building the relationship.
Click to expand...
It's a slippery concept too, because it may be hard to tell whether the propensity of the horse to stay with you is due to prior training designed to make the horse stay with you (which can be done by punishing not staying, as well as rewarding staying), or due to the horse wanting to be with you because it finds your company and/or proximity intrinsically comfortable. Actually, the distinction between these two is quite blurry! You'd have to have seen all the interactions between horse and human up to the point of noting withness.

That's why I am always a little reluctant to join in the cheers of "Look! What a fantastic relationship they have!" when e.g. commenting on some YouTube clip or other, when what we may be seeing is 'merely' excellent prior training, impressive as the end result may be.

Maybe I am being too hard-nosed about this - maybe that's what a "fantastic relationship" really is! Personally, I'd like to believe that there was some other quality giving value to the relationship other than 'merely' training. What do you think?
 
I'd like to say the pretty young Danish woman has clearly got a fantastic relationship with her horses but, having thought about it in my dreams last night, I notice she is still carrying a schooling whip, as does the Frenchman with his Highland ponies! I don't suppose either beat their horses, but I'd not rule out a bit of -ve reinforcement either!:D

I know when I train a dog to a fine degree, some of the initial ground work will sometimes include a bit of dominance (-ve). Once the ground rules are established, it only takes a hint -- as the girl's raised finger in one session demonstrated. I often describe successful training as like dealing with the Mafia. If the mob make a polite request, you comply -- in the full knowledge that to refuse may mean you will wake in the morning sharing your bed with your horse's head!

Words/names are needed when explaining things to third parties. I would handle up to 40 gundogs here, all at the same time, without using words much and seldom using names. All I had to do was make eye contact with one dog and follow that with a gesture. If you are doing things every day, animals will read body language quicker than words so it is not so difficult (although it looks impressive) and it is much the same as communication between a well tuned husband and wife team that seldom needs words or even the subtle gesture.
 
fburton said:
At best, it is a wasted opportunity in my opinion; at worst, it is letting the horse down - even though it is almost always done with good intentions.


It's a slippery concept too, because it may be hard to tell whether the propensity of the horse to stay with you is due to prior training designed to make the horse stay with you (which can be done by punishing not staying, as well as rewarding staying), or due to the horse wanting to be with you because it finds your company and/or proximity intrinsically comfortable. Actually, the distinction between these two is quite blurry! You'd have to have seen all the interactions between horse and human up to the point of noting withness.

That's why I am always a little reluctant to join in the cheers of "Look! What a fantastic relationship they have!" when e.g. commenting on some YouTube clip or other, when what we may be seeing is 'merely' excellent prior training, impressive as the end result may be.

Maybe I am being too hard-nosed about this - maybe that's what a "fantastic relationship" really is! Personally, I'd like to believe that there was some other quality giving value to the relationship other than 'merely' training. What do you think?
Click to expand...

Distance is the key to this in my view. In a confined area like a manege, it's easy to enforce the horse to stay with you as you suggest, as you are always fairly close to the horse.

However, get into an open space where the horse can put a lot of distance between you and the enforcement is less effective.

Horses, use space to protect themselves, hence the 3/4 mile flight. They also live in the 'now' which is also a consideration.

I believe that the relationship can be judged more accurately as I described previously, where bribery is not an issue.
 
Dry Rot said:
I'd like to say the pretty young Danish woman has clearly got a fantastic relationship with her horses but, having thought about it in my dreams last night, I notice she is still carrying a schooling whip, as does the Frenchman with his Highland ponies! I don't suppose either beat their horses, but I'd not rule out a bit of -ve reinforcement either!:D

I know when I train a dog to a fine degree, some of the initial ground work will sometimes include a bit of dominance (-ve). Once the ground rules are established, it only takes a hint -- as the girl's raised finger in one session demonstrated. I often describe successful training as like dealing with the Mafia. If the mob make a polite request, you comply -- in the full knowledge that to refuse may mean you will wake in the morning sharing your bed with your horse's head!

Words/names are needed when explaining things to third parties. I would handle up to 40 gundogs here, all at the same time, without using words much and seldom using names. All I had to do was make eye contact with one dog and follow that with a gesture. If you are doing things every day, animals will read body language quicker than words so it is not so difficult (although it looks impressive) and it is much the same as communication between a well tuned husband and wife team that seldom needs words or even the subtle gesture.
Click to expand...

Dogs and horses are polls apart.

Of course the horses in the clips have been trained using some -R, I really don't think that it is a realistic option to suggest otherwise. I hope at this stage we are still not using the concept of -R as a bad thing, as this was no ones intent, particularly Skinner. (Its a negative word is negative).

Carrying a schooling whip is not a problem, it's what you do with it, as always.:D
 
Last edited:
Pale Rider said:
Distance is the key to this in my view. In a confined area like a manege, it's easy to enforce the horse to stay with you as you suggest, as you are always fairly close to the horse.

However, get into an open space where the horse can put a lot of distance between you and the enforcement is less effective.
Click to expand...
Good point!

I believe that the relationship can be judged more accurately as I described previously, where bribery is not an issue.
Click to expand...
Or previous coercion...
 
Pale Rider said:
Dogs and horses are polls apart.
Click to expand...

That remark did not surprise me as it is what I am used to! I have had it before. Example, "But I am interested in training a JRT, a Labrador is totally different!" Well, they are not, nor are horses that different. Both are social animals, for a start.

I suggest you Google "comparative psychology" for some interesting reading.

Even a single celled amoeba will react to stimuli. Not sure that they can be trained (!) but I HAVE trained a few bumble bees!:D I am quite serious on this.;)

How animals react to various stimulus varies according to their adaptation to their particular niche in the environment. So most react in a similar manner when sex, hunger, extremes of temperature, threats, etc. are involved. That is why studying the behaviour of different species is so interesting. I was not surprised that my comments about falconry raised not a single remark. The behaviour of raptors appears to be poles apart to dogs and horses, mostly because they are not social animals, but many behaviours are similar -- so I've used an old technique for training a hawk to accept the hood to get a head shy horse to not only accept the head collar but to put his head into it with enthusiasm!

But, thank goodness we disagree about something, it makes for a much more interesting dialogue when we disagree occasionally!:D
 
I wasn't trying to be facetious, but I really do find training dogs very different to horses. We have working border collies which we train from pups to work sheep, and although they have some in built advantages, as do our quarterhorses, so from that respect very similar. I do find the training techniques vary widely between the two species, as far as behavioural training goes as we have discussed within this particular thread.
 
Pale Rider said:
I wasn't trying to be facetious, but I really do find training dogs very different to horses. We have working border collies which we train from pups to work sheep, and although they have some in built advantages, as do our quarterhorses, so from that respect very similar. I do find the training techniques vary widely between the two species, as far as behavioural training goes as we have discussed within this particular thread.
Click to expand...

I'm afraid my writing style is sometimes a bit abrupt. It doesn't mean anything. So don't worry about it and I'm sorry if I sounded critical.

I was at vet college briefly, then dropped out. In the first year, we did something called "comparative physiology" which involved dissecting various creatures and noting their differences and similarities with reference to evolution.

Briefly, fish have a two chambered heart, amphibians a three chambered heart, and mammals have a four chambered heart. Mammals go through all three stages during foetal development. Why should the brain not have developed with evolution too, with the same similarities and differences? I find it interesting that mammals have so many similarities in physiology yet it is assumed their brains, and behavioural responses, must be so different? Yes, there are variations but only so far as they are necessary for specialisation to exploit a particular niche in the environment. Nature is very economical and does not change things unless necessary.

I have trained a lot of dogs, of different breeds, over a very long time and I can assure you there are parallels in the training of dogs and horses. They both respond to -ve and +ve reinforcement for one! The first horse I broke, an Anglo-Arab, I treated almost exactly like a dog (because I knew no better!) and it worked! But then my methods for training dogs may not be the same as yours!:)

When I was watching the Danish girl and her horses following her around, my mind instinctively strayed to footage on YouTube of hounds following a huntsman. The principles are the same. I also use pressure and release to teach a dog to lead. I described elsewhere how I taught ponies to come to the sound of a gunshot. Gundogs must learnt to Sit to a verbal command which is then chained to other "triggers" that produce the same response -- hand signal, whistle, flushed game, etc. all result in the trained dog going into the Sit position (or ought to!). Triggers are chained in horse training too -- verbal and directional signals (with the lunge whip) on the lunge, then coupled with physical pressure on long reins, then by thigh pressure and heels when ridden. Once I've trained a dog, I expect to be able to work it with my hands in my pockets most of the time. It is almost all body language and I suspect it will be the same with you and your horses. Do you see what I am getting at?
 
Going back to the film of Marianne and her being followed by the horses, it's not difficult to do, but like dogs I suspect, there must be something 'in it' for the horse.

There are fundimental differences in the two species, and for horses the advantages of being with the human are, I believe, slightly more critical out of the two.

Whilst horses will respond well to food and attention as +R, as well as reward. These in themselves will not provide the depth of relationship we are seeking. I suggest that food for horses is mostly in abundance, not something they have to seek out and hunt down successfully, and therefore not such a strong motivator as with other species. As evidence of this, horses may be fed buckets of food several times a day, and still may ignore or positively avoid the human and be impossible to catch.

My suggestion is that to forge a meaningful relationship with a horse, this need not take more than a few moments, provided that you demonstrate to the horse some benefit that it should be with you.

Horses find leadership and safety more of a motivating factor that mear food reward. Therefore, it is essential to prove to the horse that you are a bona fide leader, able to provide mental comfort and safety. It is these motivations in my view that are not adequetly explained in Skinners theories, as they focus on the steps to getting near this position but never quite there. Probably because his models of rats and pigeons are just not sophisticated enough. Whist people are of the mind that using -R and the whole gambit, instead of exclusively +R, is some form of failure, they will never build the strength of relationship they could have as the horse finds them unconvincing and inevitably weak.
 
Pale Rider said:
Going back to the film of Marianne and her being followed by the horses, it's not difficult to do, but like dogs I suspect, there must be something 'in it' for the horse.

There are fundimental differences in the two species, and for horses the advantages of being with the human are, I believe, slightly more critical out of the two.

Whilst horses will respond well to food and attention as +R, as well as reward. These in themselves will not provide the depth of relationship we are seeking. I suggest that food for horses is mostly in abundance, not something they have to seek out and hunt down successfully, and therefore not such a strong motivator as with other species. As evidence of this, horses may be fed buckets of food several times a day, and still may ignore or positively avoid the human and be impossible to catch.

My suggestion is that to forge a meaningful relationship with a horse, this need not take more than a few moments, provided that you demonstrate to the horse some benefit that it should be with you.

Horses find leadership and safety more of a motivating factor that mear food reward. Therefore, it is essential to prove to the horse that you are a bona fide leader, able to provide mental comfort and safety. It is these motivations in my view that are not adequetly explained in Skinners theories, as they focus on the steps to getting near this position but never quite there. Probably because his models of rats and pigeons are just not sophisticated enough. Whist people are of the mind that using -R and the whole gambit, instead of exclusively +R, is some form of failure, they will never build the strength of relationship they could have as the horse finds them unconvincing and inevitably weak.
Click to expand...

I was thinking about this while attempting to catch the moles that are devastating my pastures today!

I don't pretend to be highly knowledgeable about horses but physical contact seems to be very important to them. Also, I would guess that the herd instinct is stronger in horses than the pack instinct in dogs.

One thing I would do with my young dogs was to take them onto unfamiliar territory with few distraction and deliberately "lose them". My aim was to get them to stay in contact but work at a distance. I was training working pointers which are bred and trained to range out to hunt, sometimes a few hundred yards. Young dogs can get a bit excited and will get lost as they forget to check back. Getting lost, as you've hinted, for a young insecure social animal like a dog (or young human, come to that!) is truly scary.

As I said, I trained that Anglo-Arab like a dog because I had no one to advise me to the contrary. At the time, I lived up the back of beyond in a huge block of forestry. I'd lead the horse out into the forest, making a circuit around fallen trees and through burns, until I was literally in the middle of nowhere and then remove the head collar. The horse had a choice. Stay there alone or follow me home. He always chose the latter, learning to jump trees and water. We developed quite a strong bond. He was always quite anxious not to lose me!;)

There is another technique which works for hawks, dogs, and horses. I would tid-bit whatever I was training every time I passed it. (I am currently doing this with a foal spoilt by an over-protective mum!). The hawk would get a small piece of meat, the dog a pat and a biscuit, and the horse gets a couple of slices of carrot and a scratch. Repeat that a few times each day and each is equally eager to meet you. The reactions are similar, confirming my belief that there are many similarities in training different species and it is a mistake to think in terms of just that one species you are attempting to train.

BTW, the bumble bees were trained according to a method described in a scientific magazine I had read. A bee is caught and imprisoned for a few hours, so it is hungry. It is then taught to drink sugar solution from a pipette. It is just a short step to poking the pipette through a piece of cardboard and teaching the bee to fly to a mark where it would find the pipette! Simples!;)
 
Last edited:
You must log in or register to reply here.
Top