Allow the Right, correct the wrong: a primer on praise

[Tea]

Denice

 

really good information once again. I'm with you on the making a dog wide. I am slow to do that because my dogs widen naturally on terrain and with the right sheep and with maturity.

 

Tick I used a wire basket muzzle which actually corrected him for biting. Then I showed him proper grip when he finally would take advice.

 

Old Tick- Is he nervous? You know I have seen the slash and dash dogs that are nervous and they do that out of fear. A brave dog stands his ground like a Terero and waits. But Tick it is something else I believe he throws back to a predatory rather than herding behavious.

[alligande]

This has made interesting reading and made me think about my own use of negatives. I am one of those that does talk to my dog, in conversational tones, along the lines of "what are you doing", my husband does the same thing and over the years all our dogs have "commands" that they have picked up and never been taught, they also know the sounds/words for a correction, "what are you doing" "not yours" "ahah" etc these are also used in a conversational way. I was thinking about Mr McCaigs description of his walk on one of my own we don't do a lot of leash walking but currently have to for a muscle strain and I was using verbal corrections, for things like don't look at the cat, you don't need to pee on yet another fence. This is just daily living, we expect our dogs to have good manners, at home and away from home, he can sleap under a cafe table, hang out in a hotel, be polite to children, stand in long lines to board ferries with other dogs screaming kids and drunk adults.

Which brings me to the way I teach agility, where there is no verbal negatives used, rewards are sometimes withheld, the dogs are usually rewarded when they are wrong, because they don't know they are wrong. Agility is a fast and complex sport and most importantly it is a team sport, both team members need to be correct. Dogs read our body language without any training, they intuitively understand if our feet are pointing one way that is the way to go, human error makes up a large part of the dogs errors and most often the dog was right. As an example beginer teams often miss a jump on a small sequence, and it is often not the dogs fault, the dog did what the person was showing it, so we reward the dog. Sometimes the dog is wrong and it is usually because it doesn't understand the criteria so we make the exercise easier so it can succeed then we reward, and progress.

Agility dogs need to be comfortable trying new things at speed, we don't want them thinking something might be wrong and being cautious in their approach, we want them to try again and again until we both get it right. As an example a large part of one of my intro classes is made up of people who came from another club, one that uses corrections of all type. One of these is a husky who hated being involved with agility, her very serious German owner was as stressed as her, failure was not an option. Now when they are in the ring I can't help smiling the husky has one of those great husky smiles and her owner can laugh when it is going wrong. They are never going to set the agility world on fire, but they are having fun, lyka has learned that playing with her owner is fun, and she has the confidence to try and as a bonus they have their first couple of Qs.

I am lucky enough to have a dog who finds agility with me extremely rewarding, at the end of a competition run it is our celebration that is the most important thing to him, playing tug is just a bonus, but if I stopped playing in training and competition I have no idea how long praise alone would work and it is not something I am willing to risk.

[mum24dog]

My timing of cues is terrible, always has been bad but worse since I am so rusty. If my dog responds but at the wrong time for what I intended too right he will be rewarded. Next time it might be what I want him to do and I don't want any hesitation.

 

As a general rule it is much harder to get male owners to understand the nature of reward and to put it into practice. Too many seem to have the idea that a dog should just do as it's told because man is superior and they don't get far in agility with that attitude. They can get their dogs to negotiate the obstacles but in a mechanical way rather than with enthusiasm. To be fair a few women struggle with the idea of reward too but on the whole women are more likely to get it. I'm sure that's one reason why the dog sport world has a much higher proportion of women taking part than men. Also most women don't mind making a fool of themselves by getting things wrong but men often don't accept it so readily.

[Donald McCaig]

Dear Doggers,

 

Ms. Alfreda writes (in part): "I was trying to imagine an analogy to support Donald's assertion: eg: If a complex behavior comes out of the dog in bigger chunks and whole patterns as it does in herding (gather, fetch, stalk, drive, balance, hold), then perhaps shaping that is more like sculpture? A sculptor works in negative space from a whole piece of stone and chips away at it until she refines it into the desired image. The sculptor takes *away* what she doesn’t want."

 

Interesting analogy. In a general sense, sheepdog training is progressive: one starts near and increases the distance, one teaches the fetch before the drive and the gather before the shed, but it is not systematic nor combining simple skills for a more complicated performance. It's as if the dog must understand everything (however poorly) all at once. The trainer is the captive of the dog's genetic knowing, its ignorance, one's experience (or lack thereof) and what the training sheep will let you teach. (No point teaching a young dog shedding on kneeknockers).

 

It is a living process and an opportunistic one: although shedding is usually one of the last things taught, if a good gap appears and the pup is capable of understanding it is learning something new, call him in; if a dog finds itself driving before it has a reliable fetch let it drive for a bit. If a dog lacks confidence and grips a defiant ewe, approve it.

 

Every open handler, including those who buy fully trained trial dogs must train and though the goals are (broadly) uniform the dogs are very different and one adapts one's training (and training philosophy?) for each dog. Every dog I have trained or retrained has changed me.

 

Donald McCaig

[juliepoudrier]

^^This is a good description of how most of us train a stock dog.

 

I liked Alfreda's analogy with sculpture, except that as Donald points out sometimes we try things that might not work with a particular dog (which means changing tactics midstream), and if the opportunity presents itself, we will jump ahead in the training "timeline." Both practices that could ruin a sculpture, if we're talking about sculpting in marble, for example, where there's no putting back (and where one presumably starts with a clear vision of what the finished product will look like). Especially with beginning trainers, the taking away of things does happen often enough, and sometimes those things can't be put back (or not very easily), the grip being the most obvious example, or the novice handler correcting a young dog that is clearly trying to cover bolting stock (vs. chasing them--they can look a lot alike if you're not used to reading a dog's intent).

 

Training for stock work is definitely opportunistic--I routinely will call youngsters (even barely started ones) through when gaps appear. At that point it's not even about completely a shed per se, but really about the dog coming toward me through sheep, quickly and willingly. What they learn in those brief moments will stay in their memory for the future when I finally do get around to training a real shed.

 

The other thing I routinely do is use a youngster for real work, when practical. It means planning for the chore to take more (sometimes much more) time, for mistakes to happen, etc. Like with everything in life (humans and other animals alike), I think a young dog better understands the "why" of the training if they put it to practice in a practical application. (No matter how nervewracking that can be for the trainer.)

 

Regarding praise vs. correction, I do praise and correct when training a stockdog, and even with my fully trained dogs. Where it becomes a problem (as others have said) is when the talk becomes a constant stream (or that fine line where it takes a young dog's attention away from the work, which would be more problematic with a young dog who is hesitant in its work, and which you want to encourage as much as possible). But especially with youngsters, or when a trained dog has managed a particularly difficult task, I have no hesitation in telling them they are good dogs. That doesn't mean we throw a party for a successful pen of difficult sheep, for example, but I will be sure to let the dog know that I am pleased with his/her effort that led to our success (or lack thereof).

 

Somewhere back in this long thread someone mentioned a dog who really enjoyed clapping. Be careful what you wish for! I had a little dog I bought when she was around 3. She had run in the Bluegrass as a nursery dog, been bought for someone else whom she didn't suit and then came to me at a good price because she didn't shed well. In her previous life she had been used as a demo dog at the Kentucky Horse Park and so she was used to applause and LOVED it. She would jump or run around barking in response to applause (which of course pleased the folks clapping). Normally this wouldn't be a huge issue, and it was pretty cute to see, BUT one time at the Montpelier SDT, where there are always lots of spectators (because it's held in conjunction with a fiber festival), we were running in open and the spectators where clapping for every obstacle made. At a distance on the drive it wasn't a real problem because I could quickly stop her with a whistle. but when we got to the pen and penned the sheep, the applause was loud and right there. Kat went flying around the pen tail up, head up all excited over the applause FOR HER! I had to be damn quick with the gate so that the sheep weren't pushed back out while she responded to her admiring fans. After that, we got our shed with just a couple seconds left on the clock (a miracle in itself because she was not a consistently good shedding dog) and once again the cheers and applause thundered. Kat took off up and down the fence line barking and dancing, putting on a really good show for the spectators. I, on the other hand, was just grateful that her joy over being applauded hadn't ruined our run at the multiple opportunities the crowd presented.

 

I miss that little dog.

 

J.

[Maja]

I would have loved to see that, Julie! :wub:

 

(isn't it kind of odd for people to clap at a trial?)

[Maja]

Maja wrote:

“To make a long process short .....

 

Maja, you’ve had a really interesting journey! Thank you for sharing about your dogs and the evolution in your philosophy of training! This post and your other one really had me thinking! It reminded me of the issue of intrinsic vs extrinsic motivation, which is an area of debate and research in human fields of education/psychology/business as well.

 

“There is punishment unseen, and praise unspoken….”

True! …. Your ballet teacher sounds very exacting! Very high standards/ expectations. In an atmosphere of disapproval, the slightest acknowledgment or nod, can carry so much weight. I imagine that in an atmosphere of approval, the slightest frown might carry a lot of weight too?

 

I just wanted to add that I meant to say I never really moved away from clicker, it's just that sheep herding is not by the way of clicker. If I were to train agility I would probably use clicker.

 

I had three teachers in the course of 5 years, and one was just mean and unfair. We all disliked her with a passion and had only outward respect for her. The two other teachers extremely strict and very soft spoken and very fair. I did not describe well the corrections - they were pointed out, calmly, quietly. While a look of actual disapproval was withering. And with a look of approval, we grew wings.

 

I don't know where the pressure came from, because as I wrote, actually being corrected, meant being improved upon, which was a privilege - the teacher paid attention to you. A rebuke of some sort was for erring in the routine of a known exercise. The one teacher (of the two that we liked and respected) was particularly good at delivering quiet withering tones. The worst punishment was being relegated to the "dumb barre" - it was a barre, just like any other, but it was designated to be the "Barre of Shame" and we were petrified of ending up there. The third teacher actually rarely gave any telling off, and she didn't use the "dumb barre", but going through the lesson and being ignored by her was worse than anything-in the-world . So it all reminds me a lot of border collies and some pressure techniques (which often is really laughable like wiggling plastic ribbons), and their often positive reaction to correction .

[juliepoudrier]

If it's a trial with mostly stockdog people, then you don't usually get clapping (at least not till the run is over), but trials that attract the public often do have clapping spectators. That's why Kat's love of being cheered generally wasn't a big issue (which is why it would catch me by surprise at those few trials the public did attend) .

 

J.

I would have loved to see that, Julie! :wub:

 

(isn't it kind of odd for people to clap at a trial?)

[Eileen Stein]

He said: Don’t make the command the correction.

 

He recommended using all commands in a consistent, neutral, or even upbeat way. If a correction is necessary, use another word as the correction, -as opposed to yelling “Lie Down!” or using a gruffer voice for example. In an ideal world, he said, there would be a different correction word for every different transgression.

 

This surprised me since, tone of voice is often used to convey information in all kinds of dog training, and it seemed to fly in the face of what some other handlers do. Some (perhaps Donald is one) seem to have only one correction word (ACK). Others do add corrective information (tone) into the command itself. I believe Julie Hill has a progression of “correction” that goes: ask, tell, insist. I believe this information is also carried in her whistles? Certainly it’s in her body language. (Herding experts please correct or enlighten me J).

 

Some sheepdog trainers agree with Patrick, but certainly there are others who believe in using the command as a correction, and even some who use the command as praise/reward. Derek Scrimgeour, if I'm not mistaken, is one who uses it for all three: Say "lie down" in a neutral tone, if the dog doesn't respond say it in a harsher voice, if the dog does respond say it in a softer approving voice as he is lying there. Within the generally accepted ways of training a sheepdog, there are individual differences and preferences like this. But I think tone of voice in general is huge in sheepdog training, and it is definitely conveyed in whistles.

 

Very interesting discussion -- thank you, everybody. My thanks especially to Maya for the hot-and-cold children's game analogy and to Alfreda for the marble sculptor and stonemason analogy.

[Alfreda]

Most sheep training is based on negative reinforcement. ....I've seen dogs trained with praise and dogs trained only with punishment(yelling at the dog when it is wrong, ignoring it when it is right) and I see some sad dogs (my perception based on their behaviour). I see dogs confused, unsure etc.

Pam is right. (below is my self-correction) Apologies to those who suffer from SIIS (Skinner-Induced Irritability Syndrome)

I misspoke when I said:

“I was thinking about pressure-release (in both horses and herding) and how it seems like the highly skilled trainers use the lightest amount of pressure possible. If I were to understand it in (Skinner’s) behavioral terms, the bit, reins, and leg pressure start out as aversives, and they’re released or removed as in the “negative punishment” quadrant (removing something unpleasant).”

 

I often confuse the terms negative Reinforcement (R-) and negative Punishment (P-)

 

The minus in the 4 quadrants of operant conditioning means subtracting something, and the plus means adding something. In Pressure/Release training, something unpleasant: say, tension on the horse’s rein, is released when the horse responds in the desired way. The reinforcement is in the subtraction of the rein tension.

 

Body pressure (if your dog finds it unpleasant and some dogs don’t) is reduced as the dog widens out, or responds to say, a “Lie Down.”

 

R- is confusing though, because it involves applying the aversive in the first place. The punishment (rein tension) must precede the release (the Reinforcement). For R- to work, a punishment (a pressure) must be ADDed (P+) that wasn’t there before.

 

P- would be when a condition becomes “punishing” as a result of its removal. I guess I use this a lot. Like Donald, I can teach WAIT at the door by simply closing the door if the dog rushes it. I subtracted opportunity, the dog is “punished.” Or, say, dog grips sheep, session ends. Dog is punished by loosing access to sheep.

 

Like R-, P- implies the existence of a preceding condition (in this case a REWARD) that gets subtracted.

 

Correct me if I’m wrong, most “positive” trainers I know emphasize R+, but also frequently use P- (or the removal of something rewarding).

 

Just want to add that yelling (while unpleasant) isn’t necessarily a punishment, in the Skinerian sense- unless it truly reduces the behavior- If it doesn't, that’s “nagging.” Likewise praising (while pleasant) isn’t necessarily a Reward unless it actually increases the desired behavior. Cheerleading might, or might not have that effect.

 

In the case of a dog who is "not sensitive" to body pressure, herding trainers must make pressure/proximity more "aversive" or threatening to the dog, as a first step-- is that true?

 

Or, is it possible that the dog can achieve continued (rewarding) access to sheep by learning to read and release pressure itself?

 

The first approach suggests the dog must in some way fear or dread the body pressure or stick or boogie bag. The other...?

[Alfreda]

Some sheepdog trainers agree with Patrick, but certainly there are others who believe in using the command as a correction, and even some who use the command as praise/reward. Derek Scrimgeour, if I'm not mistaken, is one who uses it for all three: Say "lie down" in a neutral tone, if the dog doesn't respond say it in a harsher voice, if the dog does respond say it in a softer approving voice as he is lying there. Within the generally accepted ways of training a sheepdog, there are individual differences and preferences like this. But I think tone of voice in general is huge in sheepdog training, and it is definitely conveyed in whistles.

Eileen- thanks for clarifying this!

 

Geonni wrote about her various corrections words no, hey! and also said she uses tone of voice within commands...Probably as humans, its a challenge to keep emotion out of our words, but it is interesting when and why trainers might wish to (try) to do so:

 

Is there much variation on this in agility/ SAR/ sports circles? Do many trainers try to keep commands as neutral as possible, and use different words to convey information like "wrong" "right" or urgency vs calm? Or are command words themselves colored with those messages?

 

I agree: interesting, wide-ranging, thought-provoking discussion!

[Laurelin]

I'm not sure I understand the question completely.

 

For agility generally I don't see people letting their dog know it was 'wrong' most the time because in general the dog is right and the handler is wrong or the handler just failed to support the position for the dog. Now in cases where the dog bites or steps way out of line then yes people will obviously get firmer/remove their dog from the course/'lie down!'/etc. For example last week we ran a line where Hank pulled around a jump twice but I didn't do anything other than reward him. It was my fault, I turned too soon and he read my body language. Just was too green for the distance I was trying to get.

 

I run my individual dogs very differently but it is just an issue of necessity and personality. My first agility dog was an elderly papillon (she was 9 when we stepped in the ring together). Soft girl, very sweet, very honest and tried hard. I did a lot of encouraging and feedback. She ran for what I was saying. She wants to know she's doing right. Even outside agility she wants that kind of feedback. 'Summer go tunnel! Yay good girl! Hurry hurry jump! Good job!' I run the dog I'm training now very tersely comparatively. Basically I only call tunnels, contacts, weaves, and otherwise shut up and tell him when to go take a full line then when to come in. And I can get loud and firm with commands. He is very fast and driven. I only have time for short commands and he doesn't need my encouragement to get around a course. If I don't get loud about contacts with him, he will not stop. So a loud 'FEET!' happens. I used to hear people get loud in the ring and think they were mad but often for the intense dogs you have to get loud to get them to hear you. If I yelled that way at my papillon she would think she was in trouble or wrong.

 

I use the word 'hurry' to speed the dog up and do that for both of them though Hank almost never needs it. I don't ask for calm in agility. So no word for that. Haha.

 

There are some clicker trainers that try to keep things very neutral all the time but that is not everyone and definitely not me.

 

I may have rambled but maybe I answered something?

[geonni banner]

I use the word 'hurry' to speed the dog up and do that for both of them though Hank almost never needs it.

That would be a problem with my guys... "Hurry up!" means go pee now, and if it's repeated after they pee, it means go poop now. I chose that as a command because that's what you say when you're standing in the pouring rain with a leashed dog that just wants to go back inside. Sigh... Apartment life.

[rushdoggie]

"The leash is just decoration".

 

I like that Kristine. When competing in Obedience I use the lightest and thinnest lead I can find where the heelwork exercise dictates that I have to use one. Leash popping or worse has never been my style.

 

My pet dog students must come with an unadorned 6 foot leash by week 3 of class, and I put slipknots in them (which would come undone if their leashes get tight) and they have to keep the slipknot in all throughout the class. Most say "oh that will never work!" yet the loop stays in 90% of the time for 95% of my students. Its amazing what happens when I put the onus on the humans to keep the dog with them and happily working. Goes to show you who needs the leash, human vs dog.

 

Having pet dog classes with untrained, unleashed dogs would be nuts, and a liability nightmare.

 

My advanced students and the classes I take myself are mostly off leash.

[Tea]

I'm bringing blood sucking fly man to your class:)

[Maja]

Derek Scrimgeour, if I'm not mistaken, is one who uses it for all three: Say "lie down" in a neutral tone, if the dog doesn't respond say it in a harsher voice, if the dog does respond say it in a softer approving voice as he is lying there.

 

Yes, that's exactly right, I only want to add: except for flanks. Derek says, you always invite a dog to do a flank. So if the flank goes wrong the dog gets a corrective 'lie down!' or 'time!' never a flank command.

On the other hadn, and this is important, once the dog lies down, it is a safe place for the dog. So it does not get reprimanded when lying down, the handler does not use aggressive body language when the dog lies down. It is a "safe place to be". That way, when all fails most dogs trained that way lie down on their own - it saves many a wreck in a dog in training. It worked like a charm with Bonnie, with Darinka of course nothing worked the way one would expect .

 

Also, he said that the dog does basically what it wants, and the handler give the dog a pth with the tone of voice. The dog chooses which path to take - nice voice, or harsh(er) voice. Again, with Bonnie it worked like a charm, she really caught on the very essence of driving by me merely altering between sweet walk-on and a very mild 'tie' (my version of Derek's 'time'). With Darinka it is beginning to work now.

[rushdoggie]

I'm bringing blood sucking fly man to your class:)

 

hahaha! theres always the 5% to keep us humble!

[Tea]

BTW Tick is not a border collie

[Maja]

BTW Tick is not a border collie

A kelpie or an aussie?

[Smalahundur]

A chihuahua?

[Maja]

A chihuahua?

:lol: Probalbly, so many of them are mean little rascals

[Tea]

He is an dwarf alligator terrier crossed on a Teddy Roosevedt cowdog

[Tea]

hehehehe- no he's hangin tree

[Maja]

See? It's partly a leopard

[geonni banner]

Does Tick look like this?