Secondary reinforcers

Get to know other Positively members here.

Moderators: emmabeth, BoardHost

Post Reply
JudyN
Posts: 7018
Joined: Tue Jul 26, 2011 1:20 pm
Location: Dorset, UK
Contact:

Secondary reinforcers

Post by JudyN »

This is something I've been thinking about for a while, and touches on something Nettle mentioned elsewhere.

I tend to reward Jasper quite heavily, often not for 'being good' but for 'not being bad' :wink: However, if we are walking past a cat, or a dog I think he may kick of at, I need to concentrate on keeping him under control - which may actually involve brute force if the trigger is overwhelming, but is usually more a case of getting him to exert self-control. Faffing about delving in my pocket for a treat, getting him to notice it when his focus is elsewhere, and trying to post the treat through his muzzle while keeping a firm grip can be enough to reduce my (physical and mental) control and flip the switch in his brain - I tend to have better results by keeping moving, with a gentle but firm hold on the lead, and saying in a light voice 'Not for you,' 'It's just a silly doggy,' or whatever. Once his focus is off the trigger I will praise him and maybe treat him, but even then, I feel that digging around for a treat and posting it through his muzzle, which will break our stride, signals that 'something has happened' whereas I would like to convey a feeling that 'nothing has happened, nothing to see, move on calmly,' though still letting him know that he done good. Does that make sense?

So, my question is - as praise ('good boy') is very often coupled with a treat, and possibly ear rubs, does saying 'good boy' in a calm but happy voice when he's fallen back in step with me after being very focused on, say, a cat but he's managed not to throw himself at it (or throw himself at me in frustration :roll: ) let him know that he did what I wanted him to do and make him more likely to exert self-control in future?
Jasper, lurcher, born December 2009
MPbandmom
Posts: 1637
Joined: Sun Jan 03, 2010 7:18 pm

Re: Secondary reinforcers

Post by MPbandmom »

Not that I am that knowledgeable, but I would say Yes, a good boy is positive reinforcement and lets the dog know they behaved appropriately. If he doesn't kick off at being told good boy and not being given a treat, then all is well.

Good Girl is one of the many "cue" words that Sirius has decided means "treat." It doesn't always mean treat, but she will look expectantly in the direction of the person who used those words. Whether she knows she did "good" or just knows that the "possible treat" cue was used, I don't really know. I like to give dogs credit for understanding more than maybe some people would give them credit for.
Grammy to Sky and Sirius, who came to live with me, stole my heart, and changed my life forever as I took over their care and learned how to be a dog owner.
User avatar
Nettle
Posts: 10753
Joined: Sun Apr 13, 2008 1:40 pm

Re: Secondary reinforcers

Post by Nettle »

One the other hand, you do get people who say 'goodboygoodboygoodboy' as a kind of litany, and that means nothing to the dog. Tone is important, but most of all, restricted use, as words mean less to dogs than tone and body language. It's entirely possible that it is the latter that dogs are reading when they enjoy a word or two of 'praise'. Try saying 'marmalade' in the same tone and with the same body language you use when really pleased with your dog, and see what happens. Mine just smiled and wagged her tail at 'cabbage'.


I agree that any words that are followed by a treat are specificaly recognised - but I do have my doubts over the value of praise-words in general.
A dog is never bad or naughty - it is simply being a dog

SET YOURSELF UP FOR SUCCESS
JudyN
Posts: 7018
Joined: Tue Jul 26, 2011 1:20 pm
Location: Dorset, UK
Contact:

Re: Secondary reinforcers

Post by JudyN »

I agree that tone is more important than the word itself - it's more likely to tap in directly to a dog's emotions. Same as in humans, really - a praise word can tell you you got the right answer, but how it is said is really what lets you know that the speaker is really pleased or not. Maybe the important thing point is whether pleasing their owners becomes rewarding for dogs (I'm not sure if pleasing the owner would then be a primary or a secondary reinforcer, as knowing your owner has a positive regard for you must help a dog feel secure).

I often wonder what others think when I call Jasper away from doing something 'naughty' and when he comes, I praise him as if he's just got into a top university - I'm sure many feel I should be telling him off for the 'naughty'.
Jasper, lurcher, born December 2009
JudyN
Posts: 7018
Joined: Tue Jul 26, 2011 1:20 pm
Location: Dorset, UK
Contact:

Re: Secondary reinforcers

Post by JudyN »

Though having observed myself this afternoon - I quite often say 'good boy' in a low-key tone. Also, when he thinks he has done something worthy of a treat (interacting nicely with another dog, ignoring another dog, not eating cow poo), he often gives me a hopeful look, asking me for a treat. He doesn't give me that look if I just say 'good boy' and he doesn't think it's a treat-worthy situation; if it was meaningful to him, I'd have expected him to have cottoned on to that and expected a treat for every 'good boy'. So maybe it doesn't mean that much to him after all.
Jasper, lurcher, born December 2009
jacksdad
Posts: 4887
Joined: Thu Dec 10, 2009 10:48 pm

Re: Secondary reinforcers

Post by jacksdad »

To help figure this out, it might help to refresh our self on reinforcers/reinforcement

The term reinforcement means to make something stronger. Reinforcer is something specific used to make something stronger.

Positive reinforcement is adding in something that makes a behavior more likely to be repeated, thus stronger, it's been reinforced

technically there is no such thing as primary or secondary reinforcers, there are just reinforcers. There is nothing wrong with using the terms primary or secondary though. Primary reinforcer refers to something the dog didn't need to be taught to like. Food would be an example of this, you don't need to teach a dog to like food. If you say secondary reinforcer this should mean the reinforcer was created, and praise would be an example.

The creation of "secondary" reinforcers is heavily dependent on the use of classical conditioning. To create praise to be a "secondary" reinforcer you would do very similar to "loading the clicker"...good boy, pause, treat...good boy...pause, treat. ideally you would not be predictable in the sequence though. after a while "good boy" takes on a conditioned emotional response, and with sufficient history you can use "good boy" as a reinforcer.

This same idea could be used to make playing ball, or any number of things into additional reinforcers.

you can even build in praise as you train.... Sit, dog sits, good boy, treat... now you have reinforced sit AND started classically conditioning good boy to be a reinforcer.

NOW...there is a catch. NOT all reinforcers are equal, and their desirability/value is first and foremost 100% determined by the dog. So even though you may have created "good boy" to be a reinforcer, it's "power" first governed by how much the dog feels it's value is and then classical conditioning history behind it.

Not all created (aka secondary) reinforcers will be useful in all situations. The only reinforcer that comes close to "universal" is food.

now, my opinion....

Yes praise can be a reinforcer, but dogs have to be taught that it is. dogs do not come knowing "good boy" is a reinforcer and not all dogs will care. If you don't intentionally create it then you run into what Nettle was saying about is it the word or the tone of voice? Further, while "good boy" (aka praise) can be a reinforcer, I have my doubts about it's power as a blanket replacement for food, and I disagree that dogs should work just for praise.

I personally see nothing wrong with building a praise word to be a "secondary" reinforcer, just stay realistic about it's power, use, value etc.
jacksdad
Posts: 4887
Joined: Thu Dec 10, 2009 10:48 pm

Re: Secondary reinforcers

Post by jacksdad »

JudyN wrote:I agree that tone is more important than the word itself - it's more likely to tap in directly to a dog's emotions.
Given my education to date on these things...If you have not actually created a specific word to be a reinforcer, then your tone/emotions are what the dog is responding to and emotions are contagious to some degree. And in this case I am not entirely sure we can be saying that the dog is actually being reinforced. But I could be wrong.
JudyN wrote:a praise word can tell you you got the right answer, but how it is said is really what lets you know that the speaker is really pleased or not.
I would strongly suspect that IF you build a praise word to be a reinforcer, then you can't use that word in a scolding tone or to be anything OTHER than reinforcement. If you want your dog to know you are not pleased, then it would only make sense you would need some other word or method. one word, two meanings would seem to me to cause confusion.
JudyN wrote:Maybe the important thing point is whether pleasing their owners becomes rewarding for dogs (I'm not sure if pleasing the owner would then be a primary or a secondary reinforcer, as knowing your owner has a positive regard for you must help a dog feel secure).
going to have to think on this. I have seen this idea presented in very wrong ways, but knowing you, I doubt you mean it in those ways. going need to think on this.
JudyN wrote:I often wonder what others think when I call Jasper away from doing something 'naughty' and when he comes, I praise him as if he's just got into a top university - I'm sure many feel I should be telling him off for the 'naughty'.
want my 2 cents... good, because your going to get it :lol: ........ you are doing the right thing. Jasper came to you. reinforce that. IF you want to get all geeky and up your training, with care you could start carefully making reinforcement for returning to you in those situations contingent on how prompt he responds. But you would need to be VERY careful to not discourage him so that he continues to leave something HIGHLY interesting, to come to "plain ol' you".
bendog
Posts: 2188
Joined: Fri Jan 07, 2011 7:42 am

Re: Secondary reinforcers

Post by bendog »

Mine all know "good girl" or "good boy" to mean a treat is *likely* to be coming i.e. they will all look at me hopefully (and even recall :lol: ) when I say it, regardless of which dog I'm talking too!

I doubt, but I don't think we would ever know without brain scans, whether they find the praise ITSELF to be a reinforcer, or whether it just predicts a treat (like a click or marker word would, but on a more variable schedule) - though I know some argue that the click itself becomes reinnforcing after a while.

I do know that I can keep Poppy with me for a while on just praise, or talking to her in a silly voice - e.g. if we see another dog and I need to recall her, and then keep her near me until they have gone past if I've forgotten/run out of treats. She'll stay with me and wag her tail hopefully at me, but after a certain amount of time, once she realises she won't be getting a treat THIS time, then praise loses it's power and she's likely to bog off and decide the other dog is more rewarding.

So I suppose praise to me is similar to a reward marker in some cases, or used as a "keep going" signal in others (i.e I will keep talking to her until the other dog has gone) but always means that there is a chance of a treat which is what gives it it's power.

I really really don't think the dogs care if I'm happy with them or not!
User avatar
Nettle
Posts: 10753
Joined: Sun Apr 13, 2008 1:40 pm

Re: Secondary reinforcers

Post by Nettle »

Some lovely analyses there :) I love you lot: you make me think!

I suspect the 'praise by itself being rewarding' thing to be a relic from old-style training where the dog is supposed to care about pleasing the owner. We are often told how 'dogs love to please'. In reality, my take is that the dog feels safest when the owner is in a good mood, so if the owner expresses 'happy' in tone, specific words, or body language, the dog may or may not link that happy state with something it has just done, may or may not care why the owner is happy, but understands that life is better when the owner is happy (unless 'happy' relates to something the dog doesn't want to happen eg having a bath, being dressed in a Santa outfit, going to a show etc).

'Superstition' is a recognised behavioural pattern in animals as well as humans, and so the dog as a species has long ago proved to scientific satisfaction :wink: that it will link certain behaviours to certain results even after a single experience, and offer those behaviours in the anticipation of the desired result.

So IME the dog wants to create an emotional environment where it can feel safe. 'Safe' probably equals 'happy'. That's rewarding in itself.
A dog is never bad or naughty - it is simply being a dog

SET YOURSELF UP FOR SUCCESS
JudyN
Posts: 7018
Joined: Tue Jul 26, 2011 1:20 pm
Location: Dorset, UK
Contact:

Re: Secondary reinforcers

Post by JudyN »

I still have a feeling we're missing something - which is that our dogs don't just want us to be happy, but they also like our 'approval', or positive regard. Otherwise, we might just as well be inanimate treat dispensers to them. And all of us here know that feeling of emotional connection with our dogs. Maybe this is just my own superstition, and it is very dangerous to read more into this connection than there is and project our feelings onto them ('Oh, Fifi just loves to dress up as Father Christmas, look at that big smile on her face').

And really, I'm not too fussed about whether 'good boy' is reinforcing or not, because that suggests that the end is getting the dog to do what we want. I see it more as part of the normal daily communication we have with our dogs, when we chat to them about what a lovely day it is, how their tea is sooooo tasty, when we say 'no, we can't go that way because the golfers are out in their droves'. The dog may not understand exactly what we mean, but seems to get the general drift surprisingly often - and regularly leaves us gasping, 'But how did you know what I meant?' So I say 'good dog', and I hope that he gets the general gist, that I approve of him.

I am extrapolating from human relationships in this (a child told to 'come here' by a parent doesn't just comply because he expects a reward), and would never want to go down the 'My dog should work for praise alone' path, but there are so many parallels between dog and human behaviour, even in the tone of their 'voice' when they are happy, sad, confused, relaxed, whatever, that I always think it is a good starting point.
Jasper, lurcher, born December 2009
User avatar
Nettle
Posts: 10753
Joined: Sun Apr 13, 2008 1:40 pm

Re: Secondary reinforcers

Post by Nettle »

Discussion includes plenty of room for differing viewpoints - I am one who thinks dogs don't give a rat's behind about our approval :wink: their concern being primarily about themselves and seeing their environment as it affects them. If they appear to like our approval, or go out of their way to cause it, that's only because the result is pleasant for the dog. But that's just my opinion.
A dog is never bad or naughty - it is simply being a dog

SET YOURSELF UP FOR SUCCESS
mansbestfriend
Posts: 301
Joined: Mon May 20, 2013 7:35 am
Location: Queensland, Australia

Re: Secondary reinforcers

Post by mansbestfriend »

Hi. I agree when a dog feels safe and thinks (as I understand it) that their person understands them, the two-way bond/respect is stronger.

I've recently had a crash course on dog behaviour *Greyhound style* :) . I'm learning they're 100% dog but definitely not 100% 'typical'. Life-rewards seem to be more effective. If he's (Max) less pushy/jumpy about going outside at walkies time, it's "good boy" followed by going outside to start walkies, less pushy/jumpy at meal time gets a "good boy" followed by putting down his meal, less barky when I walk in (after being out) gets a "good boy" followed by a ear scratch or whatever. (So secondary reinforcement followed by primary very soon afterwards.)

Sometimes the main communication is just light-hearted babble about the status quo. I'm currently reading a book called "Why It's OK to Talk to your Dog" (not exactly light reading, but interesting).

Kelly my own dog, thrives on treats and repetition during training. We have a fairly close bond I think, but my treats or tennis ball become worthless, if someone else has 'better' treats, or a 'better' tennis ball. :(
The journey of a thousand miles begins with a single Sit.
JudyN
Posts: 7018
Joined: Tue Jul 26, 2011 1:20 pm
Location: Dorset, UK
Contact:

Re: Secondary reinforcers

Post by JudyN »

Nettle wrote:If they appear to like our approval, or go out of their way to cause it, that's only because the result is pleasant for the dog. But that's just my opinion.
Of course, you could say the same about altruism in humans - it's possible to argue that there's no such thing :wink: If someone wanted to argue that I only give to charity because it makes me feel good, I don't think it's possible to prove otherwise. Similarly, if a dog does care about our approval, then by definition our approval will make him feel happier (or more secure), but it's not possible (or maybe even meaningful) to say whether he feels happy because he gets the approval or he seeks approval to make himself happy - how could you tell whether the aim was approval or happiness when approval causes happiness?
Jasper, lurcher, born December 2009
User avatar
minkee
Posts: 2034
Joined: Wed Apr 06, 2011 7:58 am
Location: Yorkshire
Contact:

Re: Secondary reinforcers

Post by minkee »

I don't think I saw this mentioned above, but there's also the theory (did I read it on here or elsewhere?) that the old fashioned trainer's dog DOES find praise rewarding.... because it's a 'no punishment' marker, like the opposite of the much discussed 'no reward marker', and so would fall into the "phew" quadrant

I can never remember my quadrants properly, but I can always remember the Homer Simpson picture demonstrating them all:

Image
emmabeth
Posts: 8894
Joined: Tue Oct 17, 2006 9:24 pm
Location: West Midlands
Contact:

Re: Secondary reinforcers

Post by emmabeth »

Glad you mentioned that Minkee, because thats one of my 'pet' theories..

Old fashioned, generally male, handlers of, generally working dogs - (lots of generalisations!), dogs spend a lot of time isolated from people when not working, handler does v little talking, no fuss, no treats... but, great big huge but..

1/ Handler = work, and the handler enables that work and work = REWARDING! Because we are of course talking collies, spaniels, working breeds whose jobs are hugely self rewarding.
2/ANY kind word from handler = relief, not going to get punished!
West Midlands based 1-2-1 Training & Behaviour Canine Consultant
Post Reply