Dear all,
I went to see an australian clinician a couple of years ago who said that he doesn't believe in +R because horses pretty much live in a negatively reinforced world and he tries to, in his words 'live in their world as much as possible'. I would welcome your views on this, Is he right, do you think that they communicate with each other using only negative reinforcement?
best wishes,
cath
I think it depends on what a horse would call positive reinforcement or a reward. I often think that just leaving a horse alone is a reward, that walking away and giving a horse space is a reward but then I suppose that is just the cessation of negative reinforcement. They certainly don't give each other things to eat (apart from mares to foals) and my lot would be hard pressed to share their food with anyone except the most beguiling Exmoor yearling. Hmmm. Need to think about this one.
They do, however, give each other companionship, mutual grooming, guidance round the field (when my mare walks to the hedge to eat and the others follow, she is using no negative reinforcement, but it could be argued that they are simply copying - as with the standard concept of the 'lead' mare who leads the herd to water/fresh pasture and the like)
Riding isn't natural. If we're going to ride our horses, we are, by definition, creating an unnatural environment. That said, it seems to me we can do our best to find the best/most efficient == least stressful means by which to communicate.
It also seems to me that Bob Bailey was right when he said that if we couldn't train a chicken to play the guitar, we shouldn't be working with dogs and horses. From a man who could and did train 16 chickens simultaneously to do far more complex things, this is probably an overstatement, but when I talk to those who have been on his 'chicken camps' they have a quite spectacular insight into the possibilities and 'rules' of operant conditioning (+R).
so - riding= unnatural. Training = unnatural, but we want to train the horse to be as efficient as possible in his/her ability to carry us and we want our training to be as efficient as possible. That doesn't necessarily mean 'natural' given where we are heading for.
Hi Manda,
Yes, I agree totally that where we're heading for isn't natural at all. I also believe that we should pick the least stressful way possible. To me, the least stressful way is the one the horse finds easiest to understand??? However, I screw myself up in knots about this one. You see, I've spoken to many people who believe that the least stressful and clearest way of training is a method that the horse recognises. However, for this to work you really have to understand just what horse culture is founded on, hence my original question. I think everyone on this site probably agrees with your definition of leadership. I know I certainly do, so for me the idea of alpha leader and domination etc. is archaic and just not worth bothering with. However, as a person who wants to learn more about +R (so I hope people don't take offence at any questions - it's purely for my education!)I worry about not giving horses clear 'no's' and only ever positively reinforcing. It seems to me that horses say 'no' to each other especially in a domestic setting. Your thoughts would be valued more than I can say.
best wishes,
cath
What can I say Cath? You should have been at the APBC Equine Day last saturday when we went through stress, learning, ethology etc and understanding how it all slots together! I haven't time to go into it all now but I hope Jenni, Donna and I will be putting together some articles etc on this very subject.
There is plenty of research out there on the impacts of different types of learning on animals, but just to get started, go and read DOn't SHoot the Dog by Karen Pryor....
Helen
Hi helen,
Actually, I was at the conference. Although i found it very interesting, I have to say that personally, it was pitched at too low a level for me. I don't mean any disrespect by that as i appreciate that you had to pitch it where the majority of people were at and for them I'm sure it was an incredible learning experience. However, having read lots including Karens book many, many times, I feel the need to bounce some of my thoughts off people like yourselves. I hope I don't ever cause any offence in doing this - It's just that I've reached a place where I feel I'm not quite sure about anything at the moment!!!
Having been around dogs, especially competitive obedience for years, I've seen a lot of things I absolutely deplore - people yanking their dogs around on check chains, hitting their dogs when they plainly didn't understand etc. However, I've also seen dogs subjected to what I consider unreasonable stress and frustration by people using CT. As you said yourself at the conference, it's not necessarily the method, but the way in which its used. I couldn't agree more. This leads me to think that possibly I just haven't seen enough good clicker trainers. There certainly seems to be a derth of top handlers who only use +R. Most of the successful ones use a combination of +R, -R and even +P. You said yourself that probably, this is the fastest way to train something. This is where I get stuck - one of the central themes in CT is ignoring 'unwanted' behaviour. I really struggle with this one especially with regard to dogs. This is what made me ask the original question - what do horses do?? I don't think they ignore each other in the way that we would if sticking to this central dogma of CT.
respectfully,
Cath
I may be out on a limb on this one, or re-inventing a square wheel, or whatever, but did anyone ever say that we can't say 'no' when we use +R operant conditioning?
Alex doesn't and she's the one I learned from for horses. Certainly none of the dog people I ever learned from did - and I have an 'untrainable' lurcher who hasn't had a collar or lead on for the past 5 years or so who is completely relaxed, safe and sane who was trained with a mixture of clicker training, and a solid 'I really don't think you meant to do that' (abbreviation 'NO!') which seemed to me to work pretty well
so in working with Banner, who is my sole personal experience of clicking with horses and therefore hardly representative, I'm using a tongue click paired to a reinforcera solid 'Don't even think about it. (or 'Ah', or 'No!', depending on how far she's got down the 'thinking of moving' road) and pressure/release with a head-collar with two lead reins attached.
(got the bitless bridle as recommended by Catherine and Banner hates it - webbing is too narrow and the holes for the buckles are scratchy on the flat skin beside/below her eye - anyone who wants to buy it is welcome). So we have +R, -R and a verbal +P.
She is confused fairly often, but that's because I'm crap at thinking up ways to communicate that are clear enough - and it's improving by the day. I'm heading for softness, lightness and what Mark Russell describes as 'responding rather than reacting' which seems to me a fair goal.
and
marginally off topic, Everyone but everyone says that the click 'marks and so ends the behaviour' EXCEPT Bob Bailey who was pretty clear that the reinforcement 'ends the behaviour' and anything exhibited (or required) between click and reinforcer is also reinforced. So I ask for a midline head, slightly up and in, so that, with time, she is beginning to lift her wither before she gets the miniscule fragment of carrot. And her ears need to be forward, but 'ears forward' was the first thing we clicked for and that's pretty much solid by now.
does this help? Am I teaching a schoolroom of grannies to such eggs?
Hi Manda,
Thankyou for the explanation of how you train. Actually, this is pretty much what I do too.
I was interested in what you said about Alex having a 'no'. I spent two hours in a pub just in a one on one conversation with Alex just last year. (She took pity on me cos I think she could see I have too many thoughts and am naturally a worry wort about all this stuff!!) Anyway, in answer to my question ' if we give a clear 'yes' shouldn't we give an equally clear 'no'?' Alex seemed very clear that she doesn't have a 'no'. She was also pretty clear that Karen Pryor really did mean ignore the unwanted stuff. I think she knows Karen personally. Of course, Alex may be someone who only gives people extra information when they';re ready for it, I don't know her that well yet having only done two clinics with her. Is it possible that she introduces 'no' only when people are further down the line?? However, on the more general point, I did think that simply ignoring unwanted behaviour is central to CT. However, there are people on this forum more qualified to comment on this than me.
cheers,
Cath
Forgive me if I’m wrong, it could just be your wording, but I wonder whether there is some misunderstanding. An often used example of negative reinforcement is in the area of leg aids – we apply pressure to the horse’s side until he moves. The removal (or cessation) of the pressure reinforces the moving behaviour. So it’s not negative reinforcement which ceases, it’s the pressure that ceases which equals negative reinforcement.
I have the distinct feeling I’m not making myself very clear – somebody help me out here!
Re your Australian clinician – surely, the point is that the horses we train live in our world
My take on things is that what matters is we communicate with them in a way that they can understand and are willing (not just left with little choice) to work with. I don’t believe we can do that by attempting to mimic them or their relationships with each other. I believe we can do it by taking a predominantly positive and rewarding approach, looking first at ourselves when things go wrong, thinking from their perspective and putting their interests first. We have, after all, brought them to our world to service our own desires and needs, you don’t see them queuing up to come off the prairie for a life of domestication with humans
By trying to imitate horses in order to communicate with them, aren't we seriously underestimating their intelligence and ability to learn about us? They do learn to understand what our behaviour means to them. How could we teach them ‘bad’ behaviours inadvertently if they relied on us trying to speak their language for them to learn?
I fear I may be rambling on a bit now, should have gone to bed ages ago.
Hi Jan,
You're not rambling - this is all good stuff. Would you mind expanding a bit on your last paragraph though as I didn't quite understand your last sentence.
cheers,
Cath
I’m sorry this is such a long-winded reply, but I didn’t want to risk being ambiguous. So I apologise in advance if I haven’t succeeded in avoiding that!
I’m sure you won’t offend anyone by asking such searching questions – it’s a sign of a healthy attitude towards learning and training!
This argument from the Australian trainer is one I’ve heard on numerous occasions. While it’s commendable that such people try to approach training from the point of view of the horse, I think it suggests that they haven’t thought the matter through properly and also that they don’t really understand the nature of reinforcements. I wonder, for example, what this trainer regards as the ‘negative reinforcements’ in a horse’s world? One example often used is that of the horse (A) who kicks out (or threatens to kick) when another horse (B) approaches too closely. If B then moves away and A stops kicking out, then B (hopefully) learns that if he approaches A too closely he (or she) will kick out or threaten to kick, but if he then moves away the threat will cease. In this case the kick (or threat of a kick) is the aversive; the cessation of the threat acts as the reinforcer. So this is a case of negative reinforcement (although if A actually carried on kicking and made contact, that could be considered a punisher: B is being punished for approaching too close). However, under truly natural conditions (e.g. among feral horses allowed to get on with their lives without human interference) such aggressive actions form only a tiny part of day-to-day equine interactions (they are more common in domestic situations because of the conditions most horses are kept in). There are plenty of examples of +R; Manda’s example of mutual grooming is a good one. Say one horse (C) wants to groom with another (D). He approaches D using body language which suggests he wants to groom. If D feels like reciprocating, he will stop whatever he is doing and allow C to indicate where on his body he wants to be groomed. The fact that D responds in this way is a reinforcer: C’s approach to D in that specific manner has been successful so in future he will be able to approach D in the same manner & hopefully start a mutual grooming session. Of course it may not be s straightforward as this: C also has to learn to read D’s body language to see if he is likely to be in the mood to groom. Nevertheless, the fact that his approach has been successful encourages him to try again: in other words it has been reinforced. Another example might be the foal who wants to take milk while his dam is walking. The foal nips in front of her and blocks her movement; she stops, and he moves to the udder and starts suckling. His blocking movement has succeeded: he has got his milk, and this reinforces the behaviour which led to it. And this is surely positive reinforcement. There are many more examples I could use but they would take up too much room!
I think the problem with so many trainers who come up with this kind of objection to +R is that, while they may have spent some considerable time observing horse behaviour (at least, this is what most of them claim), I suspect that in many cases this observation may have been rather superficial. Why else would so many of them be fixated on aggressive behaviour (which is easily observable because of its nature) and fail to notice that under natural conditions this behaviour is infrequent and that most equine interactions are co-operative and/or affiliative?
Regardless of all this, I think the whole issue is something of a red herring. All animals learn in much the same ways, and the principles of conditioning and reinforcements are universal. (Karen Pryor talks of them as if they were laws of nature, rather like those of physics; I wouldn’t go that far because animals can resist conditioning, but she is not so far off the mark) And all animals, regardless of species or lifestyle, respond to +R. Horses may not give each other presents of food but that doesn’t mean they can’t take great delight in being given a titbit (if they did not, we would have no problems with food-induced nipping!). And if they like and want something, whether it is food, a stroke, or something else depending on the individual and the circumstances, then it will be an effective reinforcer.
Like Jan, I don’t think it’s either necessary or desirable to try to mimic horses or their relationships with each other. They know we are not just funny-looking horses and we don’t have the right physical equipment to mimic their subtle body-language. Jan also says, ‘By trying to imitate horses in order to communicate with them, aren't we seriously underestimating their intelligence and ability to learn about us?’ I think she is right here, too. I am all for making things easy for horses, but if a goldfish can respond to +R, horses should have no problems with it!
I think what Jan was trying to say in her last sentence (and correct me if I’m wrong, Jan) is that if horses rely on us trying to speak their language in order to learn, then it would not be so easy for us to teach them the wrong thing by mistake. For example, horses can very easily be taught to kick out when learning to tolerate having their feet picked up. It’s natural for them to try to snatch the foot away if the handler holds on to it for too long. The handler needs to be able to sense when the horse is about to pull his foot away, and put the foot down before the horse does pull it away. The length of time the handler can hold the foot must be increased very gradually so the horse comes to tolerate it. However, if the handler is not on the ball and hangs on to the foot too long, the horse may then start to kick out. If the handler then lets go, the horse learns that all he has to do to avoid picking his feet up is to kick out. (the alternative of course is to hang on, but this might not be an option with some horses! In any case, we want to avoid conflict; it’s far better not to let things get to that stage). Now it’s obvious that horses do not have any body language among themselves that could possibly signify, ‘I am going to pick your foot up’, yet the horse has learned something undesirable in one easy lesson!
With regard to the concept of saying ‘no!’ then I would agree there are times when we have to say ‘no!’ However, unless the behaviour is potentially dangerous then I think we are better off ignoring it, because here again there is the potential to teach the horse something undesirable without meaning to. Some horses, like some (most?) children, can learn to use unwanted behaviour as a distraction, so sometimes taking the time to say ‘no!’ can actually act as a reinforcer. I think we have to base or judgement here on the circumstances prevailing at the time.
I believe we can use mainly +R with horses, but the nature of what we do with them means that –R is sometimes preferable or even (as in the case of correct ridden work, unavoidable) as long as it is used mildlyand does not become oppressive. There is even a place for +P – again, as long as this is used very sparingly, only when it is really required, and the timing is spot on. As for CT – this is a training tool which, like any tool, can be used well or badly. The idea of there being any dogma associated with its use is rather concerning, because as soon as dogma creeps in, independent thought flies out of the window…however as long as there are people who continue to ask questions in spite of dogma, there is hope…
So I really wouldn’t worry too much, Cath – keep on asking the questions!
Dear Leslie,
Thank you so much for that very clear reply. I very much appreciate the time you must have taken over it.
In asking the question I was aware that there are as many bad ethologists out there as there are bad trainers so I agree with you totally that often its the aggressive behaviour that people notice most. From my limited reading (its expanding rapidly!!) I also have come to learn that in feral herds aggression is not the norm and horse society seems to operate on the basis of cooperativity more than anything else. However, I get stuck with the argument that we are not dealing with feral herds and we have to deal with the horse as it is. To my mind, if the horse really was uncooperative, we wouldn't have domesticated it in the first place but I still have trouble when people say 'well look at what they do to each other in the field'!!
If you don't mind, would you have time to give some examples of when +p MIGHT be appropriate. I'm in a place at the moment where I try to make a bargain with my horses that goes something like ' I won't ever hurt you unless you try to hurt me first'. Obviously, this relies on me being sable to correctly read what they're saying. When observing what other people do, it seems to me that nerves and fear seem to be the main reason people resort to +p, or not truly understanding their horses intentions in the first place. However, at this point in time I just don't have enough other tools to cope with certain behaviours and I just have the feeling that not cancelling out +p completely might even keep me alive with some horses in some situations.
Best wishes,
Cath
p.s. thank you for encouraging me to question, but you only have yourself to blame now if I drive you mad!!!!!!!
I'm showing my total ignorance here, so do feel free to just refer me to some relevant reading material, but with my mare (owned, backed, trained and competed for seven years by a BHS instructor until this year when I bought her) if I ignore bad behaviour (i.e. nudging me with her nose when she is impatient and wants me to hurry up with my faffing and take her out to the field, or get out on a ride), she does it again and again, increasing the force she uses until she finally gets what she wants and we set off. Positive reinforcement in between (sometimes difficult to find long enough windows of opportunity) doesn't stop her continuing to nudge, although I do try and wait until she has behaved politely before giving her what she wants and going out. The next day she does it again, but more so, and so on. Unfortunately, there are things I have to do in our routine (i.e. keep her in at night in winter, and give her an inhaler in summer) which require me to delay her. If I say "no" to let her know that the nudging is not required, and also reward her positively when she is behaving politely, she often desists long enough to have some good opportunities for more reward, and over time, the nudging gets less. It's fiendishly difficult to get my head around, as it is also of course dependent on her mood at the time, if the other mares are already out, if another horse walks past, etc. and if she's in season, she's twenty times more impatient than usual and sometimes quite unhandlable when she's taken out of the field away from her mares. She's an intelligent mare, dominant in her group, with an inquisitive and friendly manner who loves being out and about and learning new things, and has bags of confidence, but will always push to see what she can get away with. (Wakes me up in the morning, I can tell you!)
Cath - It's a while since I trained with Alex and doubtless she's moved on light years beyond where I was. I doubt very much indeed if she stores things up for later in the line so it must be the case that she doesn't use a 'no'. For myself, I think it makes things clearer, particularly in the case of 'stand still when I'm doing up the girth and you think it means to move off' when 'uhuh' said quietly gives enough of an interruption to rock her back on her feet so I can C/R for standing still. I don't think I could train myself not to (! - clearly I could, but I'd have to believe it was worth it), but will look at it this week.
Thinking of the first post, the Australian trainer saying he lived in the horses world (or something similar). I might have the wrong end of the debate, but have just read something that seemed relevant. I'm halfway through Richard Dawkins new book "The God Delusion" and he quotes Harvard geneticist Richard Lewontin "the one point which I think all evolutionists are agreed upon is that it is virtually impossible to do a better job than an organism is doing in its own environment".
I thought this demonstrated that any "domestic" arrangements we make in order to have our horses, by neccesity results in the horse having to make compromises. Some will have to compromise more than others, and some will cope better with making those compromises.The horse is adapting to or responding to our world more than we are to theirs.
Great question and I agree with Lesley that it's always good to question and we won't get sick of it. Afterall, when people don't question this site goes really quiet and then I have to think of spontaneous things to say, just to positively reinforce you all for visiting!
I completely agree with Manda and Lesley, that if you look more closely at horse behaviour it is easy to find many examples of positive reinforcement (and yes, Sarah, leaving the horse alone is -R - not the cessation of -R, it is the -R itself. Of course, if the horse wants your company it could be -P....). As I said on another thread, the people who claim that horses only use +P/-R generally have a vested interest (and perhaps also a genuine belief, I'm not suggesting it is insincere) in maintaining this viewpoint. But once you have developed a method based on "-R/+P" it is much easier to sell it to the public if you call it "living in their world" rather than "applying and removing aversives". Sorry, I'm being cynical and facetious again!! It's not just about what horses give to each other, there are plenty of ways of obtaining +R - eg browsing hedgerows etc etc So when you are using +R for training the horse is just working out what he needs to do in order to obtain the reward.
Cath, I agree that the least stressful way is whatever the horse finds easiest to understand and that in real life you have to be very careful about just assuming that +R is stress-free. A horse who has never been allowed to make a decision will certainly find life a lot easier if he is just given a set of cues that he knows how to respond to. Those cues may have been taught with -R originally but sufficiently mildly that there is no lasting trauma associated with them. If you then suddenly decide to free-shape his behaviour he may well find it extremely stressful and difficult not knowing what to do. Having said that, depending on the owner, it may be much more beneficial to the horse to do some very careful and basic free-shaping work in order to teach this horse to make decisions and take some control back over his life. This is back to "feel" and the long-term best interests of the horse, whilst simultaneously remembering that horses live in the moment and won't necessarily care about their long-term interests. This is all very much my opinion as opposed to facts but if you're interested you may like to read "On Becoming a Person" by person-centred counsellor Carl Rogers (or see the article I wrote based on the book - it's on the articles page of my website http://www.equinemindandbody.co.uk ). All about personal growth in humans and the relationship the counsellor needs to build with the client in order to maximise the client's psychological health. I think it can be really relevant to horses and our relationship with them too. It is this personal growth aspect that I am most interested in when it comes to CT and so this is what I'm getting at when we've talked privately about free-shaping. For me it's not about training behaviours per se which is why I have no knowledge as to whether it is possible to produce a top-level obedience competition dog using CT, it's just not what I'd choose to do (although that's not to say I disapprove of others doing it - if I had a collie who appeared to enjoy the challenge then I would think again).
As for saying "no", I really don't have a problem with this either, UNLESS I have a situation such as the one above. That horse who is just learning to make his first ever decision is going to be very worried (initially) about getting the wrong answer. But relatively quickly (if I am competant) he will learn that there is no wrong answer and that he is allowed to express himself. That takes a huge amount of trust on his part. If I suddenly say NO then I consider that to be an abuse of his trust. I have said to him "I know this is really difficult for you and against everything you believe but I want you to make a decision in order to enhance your personality. There is no wrong answer....... Actually I've just changed my mind, that was wrong." In behavioural terminology (although bearing in mind this is actually on a much more emotional level than behaviourism allows for) I have just reinforced the peak of an extinction burst. The horse's long-held belief was that "getting the wrong behaviour will result in something bad so therefore he shouldn't place his trust in me". By reinforcing that view when it is on such a stretched out variable schedule is something I consider emotional abuse. Even just accidentally being late with the reinforcement could be enough to trigger this sort of feeling in the horse which is why I often recommend people don't CT their horses.
Obviously whether or not this is the case very much depends on the horse. I am not generalising to all horses here, just specific cases in which I would not want to have a "no" (or at least, not inside a well-defined free-shaping session). With a well-adjusted horse I don't think there is anything wrong with saying no. Provided of course you have always looked at the reasons for any unwanted behaviours and this is something that Emma talked about on the extinction burst thread so I won't repeat it all.
But why do we want this "no"? Is it a punisher or just something to interrupt the behaviour so that we can go away and plan how to shape the behaviour so that it doesn't arise again the future? I have no problem with the latter, it's how I keep safe when I've misread a situation or have to interact with a horse who I'm not responsible for so can't train. But I do have a problem with routinely using it as a punisher which is in my opinion simply a poor replacement for proper training. I'm not perfect and I don't want to make out that I get it right all the time but if a horse is displaying a behaviour that I regularly need to say "no" to then I need to think about how to train it properly and in such a way that I can say yes instead. This is all about setting boundaries with +R. It's not easy but if you think about a basic example of mugging, it is fairly basic to teach a horse not to mug you just by teaching it an alternative behaviour of touching a target and allowing it to work through the mugging extinction burst. I gave a few more examples of how I have set boundaries (and in another case said "no") in the thread "Leadership is this relevant with horses", currently on page 2 of the forum. You'll probably do more of this on your workshop with Ben.
I'm surprised Alex said she didn't have a "no" - she seemed to be using quite a lot of "no"s when I saw her. It was a demo aimed at novices so if it's something she only teaches people further down the line then there seems to be a bit of inconsistency there. Or maybe she's just changed her approach as this is going back a while now...
Hi Jenni
Welcome to the THF, it's great to see you posting on here. It's hard to say why your mare continues to nudge you but from the way you say "she does it again and again, increasing the force she uses until she finally gets what she wants and we set off" it does sound as if you are somehow still reinforcing her behaviour. And by trying to delay giving her what she wants you may actually be strengthening the variable schedule (afterall, the behaviour increases) and/or just rewarding what has now become a behaviour chain. So while in real life we often have to accept that we all have to do things we don't want to do, such as wait for our inhalers, you might find it is easier for both of you if you can work out how you are maintaining this reinforcement (eg get someone to video you?). It may just be the fact that ultimately you put her out each day and that that she thinks she needs to nudge you in order to ensure that you do. In the imaginary scenario I have in my head (which may be utterly unsuitable for your situation, it's only an example) I might be inclined to go back to basics and try putting her out without the delay so she has the opportunity to learn that she doesn't need to nudge me. When she has started to learn that, I would consider her ready to move up a stage in that sometimes she needs to wait a little bit but that she still doesn't need to nudge me. Gradually I would teach her that patience can be extremely beneficial to her and gets her what she wants. Depending of course on why she gets impatient/agitated/whatever like this - it is about seeing the world through her eyes and trying to find a solution that suits her. Have you read Ben Hart's ebook on shaping? That might help you devise a plan to get to the bottom of this - see http://www.hartshorsemanship.com
Catherine
PS Manda, I'd agree that the click doesn't necessarily end the behaviour. It all depends on the horse's perception of exactly what the bridge is. The ultimate terminal bridge is the hand going into the treat bag. Often the click is simply an intermediate bridge (or keep going signal if you like), either because the trainer has deliberately taught this (as I have) and uses treatless clicks or because that is just how the horse perceives things. It depends. Sorry to hear about the bridle - you're the first person I've recommended it to who didn't like it. Hey ho, you win some you lose some!
Thanks Catherine. I will have a read of Ben Hart's book. I think the nudging grew out of evasion of the inhaler when she was very wheezy and had to have a large dose which took a long time - she started off by hiding her nose either behind the door or behind the corner so that I couldn't get the inhaler near it, then graduated to nudging as an effective way to dislodge it. Now it has grown into a general attention demanding / hurry up signal. Horses are so resourceful! The perfect situation has arisen to go back to basics as you suggest, as her breathing has improved so much I don't need to use the inhaler at present. I'll read on it and work out how I can reintroduce without reinforcing the behaviour.
Hi there, just like to comment on Catherine's post below.
[PS Manda, I'd agree that the click doesn't necessarily end the behaviour. It all depends on the horse's perception of exactly what the bridge is. The ultimate terminal bridge is the hand going into the treat bag. Often the click is simply an intermediate bridge (or keep going signal if you like), either because the trainer has deliberately taught this (as I have) and uses treatless clicks or because that is just how the horse perceives things. It depends. Sorry to hear about the bridle - you're the first person I've recommended it to who didn't like it. Hey ho, you win some you lose some! ]
I use Bridge and Target Training because I find the use of a continual verbal IB really supports the animal when training and then a crisp verbal TB when the behaviour is finished leaves no doubt in the animal's mind that he has understood in what you are asking. One problem with using a click as a keep going signal is that unless you constantly click it would be easy to reinforce an inwanted behaviour. Also the terminal bridge of hand going into treat bag still leaves a gap where things could go wrong and be inadvertantly reinforced.
I watched Ben do a clicker demo and he threw a cone for the horse to touch, it was a bit too far away so horse looked at cone but didn't move toward it. If he had started a vocal, continous IB when the horse looked, then the horse would have realised that he was on the right track and possibly taken a step towards the cone, he could have then intensified the IB to encourage movement and support the horse in making a decision to move to the cone and touch it. As the horse would have been focused on the cone the verbal TB when he touched it would end the behaviour more precisely than the horse eventually noticing the hand going into the treat bag and returning to him.
This method also reduces the frustration sometimes experienced by the animal in trying to work out what to do next.
It also makes training so much faster as the animal is instantly informed if he isn't on the right track and continually supported when he is doing the right thing. Very complex behaviours can be trained in sometimes record time.
We trained a goat at Wood Green Animal Shelter to target a hand then sent him away to a remote target (which he had to climb into a trailer to reach, negotiating obstacles along the way, and then return to hand target in record time using this method. He was bridged enthusiastically all the time and really looked like he found it all great fun.
This goat had been sent to Wood Green because he used his horns on people and was deemed dangerous. It was virtually immpossible to put a head collar on him but again using this method he was soon putting his head into the head collar and standing calmly whilst it was done up. We had to use some time out initially so he could work out that trying to impale people wasn't going to earn him any rewards. I don't have a problem with this and I don't think he did either.
We did find that when we produced a crop to use as a target stick he reacted by shaking and dropped to the ground in fear, so it would be fair to surmise he'd been beaten in the past.
Sorry I've gone off the point a bit!
Warm regards, Sue
Hi Catherine and Manda,
on the subject of 'no's', Alex Kurland did tell me that she experimented with using 'no' a few years ago (maybe this is when you saw her??). However, her conclusion was that although 'no' clearly worked to reduce behaviours she didn't want, it always ended up with some side-effects too so she abandoned it in the end.
I like Alex vey much as a person and teacher. Although she is now 'no-less' for want of a better expression (!!!), she never tells people to abandon what they already know as it might just keep them safe until such a time as they have more suitable tools. However, working with her and people like her has certainly made me take a step back and look at alternatives to the ways I've done things so far. AS a few people have already said on this thread, its good to keep questioning. To my mind, if I reach a stage where I think I can't learn anymore it will be time to hang up my dog lead and riding hat!!
best wishes,
cath
Hi everyone , very interesting thread to read.
Catherine , I liked when you said
( A horse who has never been allowed to make a decision will certainly find life a lot easier if he is just given a set of cues that he knows how to respond to. Those cues may have been taught with -R originally but sufficiently mildly that there is no lasting trauma associated with them. If you then suddenly decide to free-shape his behaviour he may well find it extremely stressful and difficult not knowing what to do. Having said that, depending on the owner, it may be much more beneficial to the horse to do some very careful and basic free-shaping work in order to teach this horse to make decisions and take some control back over his life. This is back to "feel" and the long-term best interests of the horse, whilst simultaneously remembering that horses live in the moment and won't necessarily care about their long-term interests.)
I have just bought a new horse , had him about 5 weeks now.He is a 5 yrold Luso who was gelded in April after coming over from Portugal. It is obvious by his behaviour he has been very regimented in his life.He can do almost any lateral work you ask of him but not loosely and fluidily.If you take him in the school with no side reins on ect he tries to attack you.When he came to me he had never been turned out with other horses and had little concept of what to do around them.After coming to me it must have seemed to him that his whole world had changed.He was turned out with 7 others after carefull introduction period , he was allowed to have a voice and I started to do clicker and feed him from my hand ....After 2 weeks of this he started to turn into a 'monster'. I truely believe it was because too much had changed for him , all the rules he had known and lived by all his life were out the window.So I had to take a step back and rethink my training.I decided at this moment in time not to do clicker or any hand feeding as it just really stresses him out.I will introduce it in the future but for now I think he is best left just to gradually get used to his new found freedom in levels that he can cope with.When we have develped a good bond and we know each other better , then maybe then I will step up his training and teach him how fun it can be.
Just to pick up on something Cath, I certainly didn't mean when we played the training game for people to think that they should be using a combination of positive punishment and positive reinforcement when training their horses. I do not agree with this AT ALL, as Catherine says it is very stressful for the horse and in my opinion highly unethical. However I would at times use a small amount of negative reinforcement depending on what it is I am training, the character of the horse I am working with and their history etc etc. I don't generally find it is necessary to use a 'no' signal at all when I am doing a training session training a behaviour, generally ignoring everything you don't want and rewarding what you do is sufficient. However in real life there are times when I have to use a 'no' as a positive punisher, for example in a dangerous situation where someone horse or person is about to get injured. I have a 7 month old foal that I bred this year and I have trained her to do everything using positive reinforcement and a small amount of negative reinforcement- at first just scratching and then clicker and scratching and then clicker and food, and scratching depending what we were doing, and mild pressure to ask her to step back out of my space, coupled with a positive reinforcer for doing so. However the first (and only time) that she nipped me and kicked me ( both in play, on two seperate occasions) I chased her away and shouted at her. She was shocked and moved away from me about 10 metres. I then approached her quietly and gave her a scratch for standing quietly and calmly and behaving well. This made it clear to her that the action I took was BECAUSE of what she had just done, and not because I didn't like her anymore or had turned into a raving monster. Ever since this (which was when she was about 6-8 weeks old) she has behaved impeccably with feet and teeth and is an absolute pleasure to work around. She is a confident foal, not at all scared of me and willing to try anything, yet at the same time she is gentle and not at all pushy. So I do believe that there are occasions when positive punishment is necessary. I have also used negative punishment with a horse I have worked for a client a few years ago, this is also known as 'time out' in other words removal of self, food and clicker during a training session because of unacceptable behaviour from the horse, in this instance a very deliberate and nasty bite out of the blue because the mare had never had to work for anything in her life and got very frustrated when I witheld the click to stretch her behaviour. Ok in that instance I learned from it because I was probably at fault in my training, I changed what I was doing with her and learned for the future! But it was necessary to make sure that her behaviour had a consequence that was more meaningful to her than just ignoring it, because her owner said she had done things like this in the past and had the potential to be quite a dangerous horse.
But this just gets back to the point I was making on the other thread (APBC one) that we need to understand how all the tools work (ie. positive and negative reinforcement and positive and negative punishment, habituation, desensitisation, counterconditioning, flooding, extinction) and we need to understand their impact on the horse in terms of stress etc in order to make educated decisions about what is approprate to use and when. That was what I hoped you would get out of the day Cath, and I am disappointed if this was not the case. If you have any other questions don't hesitate to ask!
I don't know that i need to say much more because I think Lesley has done it all for me, thanks Lesley!
Helen
Sorry posts crossed, didn't see your latest Cath, I agree with you we are all always learning, every horse and client and person I meet teaches me something and I hope I always have that attitude!
Rachel your poor poor horse isn't he lucky to have ended up with you now. I think by the sound of what you are saying you have done the right thing, take things slowly and introduce as muany primary reinforcers into your work with him as you can, while still working within the conventional framework that he understands- he still needs the boundaries etc that he is used to... and remembre that safety must ALWAYS come first for you and your horse and those around you... give time for the management changes to help him and for him to relax with you and learn that you are not going to abuse him but that you are predictable and do have boundaries. Once he realises that his work can also have a positive consequence hopefully things will get easier, then maybe you can look at reintroducing the clicker training but very gently and teaching him stuff that focuses him like target training etc then introducing relaxation exercises and the concept of dealing with frustration. If you are concerned about it maybe you should get someone professional involved to help you out, though I know you have a good bit of experience yourself. Hard to be more specific with the info you have given but hope this helps you feel better about what you are doing!
This message has been edited by DrHSpence on Dec 5, 2006 5:53 PM
Hi Helen,
No, don't worry, I didn't get the wrong end of the stick when you talked about -R, +P etc in training. I did realise you were talking about people. However, I actually thought that this raised another point. You see, whenever people talk about using 'no', its sometimes hard not to sound as if you beat the horse with a big stick!!! I think there are ways of using 'no' that don't in any way equate to this and don't cause stress. I actually thought that the training game you demonstrated possibly highlighted this point although I don't think it's what you intended. It was unfortunate that the lady who volunteered to be the trainer had such a nice, quiet 'no'. You see to me, the level of 'no' that she used looked a bit like that game you play as a kid where you hide something and then say 'warmer', 'hot' 'cold' to the person doing the finding. To me, this is what her 'no' was like. I actually think there is an equivalent with animals and I think it helps to make things clear to them. It's certainly the attitude I try to have when I train. The horse/dog has a puzzle to solve and I will help by saying hot, warmer and sometimes 'cold'.
There is also the 'no' you talked about above used in rare circumstances as a punisher. I totally agree that these should be rare instances and dependent on the horse and the situation.
The path I'm on with all this stuff now is getting very complicated and taking me to places I never thought I'd explore. Catherine, I totally agreed with you when you talked about the stuff on personal growth etc. You see, if you think about it, 'no', 'yes' whatever, only counts if the animal understands your intent etc. You can do all the +R in the world but if the animal feels stressed, intimidated, dominated, it's no better than using -R or +P. So, my quest is to find ways better ways of presenting myself. This has taken me quite far into learning about energy through energy medicine type stuff and martial arts. It's seems strange to me that I've ended up on this path when the place I started was with a horse I just wanted to be able to ride!! Ain't life strange!!
Hi Helen,
No, don't worry, I didn't get the wrong end of the stick when you talked about -R, +P etc in training. I did realise you were talking about people. However, I actually thought that this raised another point. You see, whenever people talk about using 'no', its sometimes hard not to sound as if you beat the horse with a big stick!!! I think there are ways of using 'no' that don't in any way equate to this and don't cause stress. I actually thought that the training game you demonstrated possibly highlighted this point although I don't think it's what you intended. It was unfortunate that the lady who volunteered to be the trainer had such a nice, quiet 'no'. You see to me, the level of 'no' that she used looked a bit like that game you play as a kid where you hide something and then say 'warmer', 'hot' 'cold' to the person doing the finding. To me, this is what her 'no' was like. I actually think there is an equivalent with animals and I think it helps to make things clear to them. It's certainly the attitude I try to have when I train. The horse/dog has a puzzle to solve and I will help by saying hot, warmer and sometimes 'cold'.
There is also the 'no' you talked about above used in rare circumstances as a punisher. I totally agree that these should be rare instances and dependent on the horse and the situation.
The path I'm on with all this stuff now is getting very complicated and taking me to places I never thought I'd explore. Catherine, I totally agreed with you when you talked about the stuff on personal growth etc. You see, if you think about it, 'no', 'yes' whatever, only counts if the animal understands your intent etc. You can do all the +R in the world but if the animal feels stressed, intimidated, dominated, it's no better than using -R or +P. So, my quest is to find ways better ways of presenting myself. This has taken me quite far into learning about energy through energy medicine type stuff and martial arts. It's seems strange to me that I've ended up on this path when the place I started was with a horse I just wanted to be able to ride!! Ain't life strange!!
Been immersed in the book for a bit and not reading the list... also thinking about Alex and her use of 'no' or otherwise.
forgetting the things we say in clinics, which may change, one of the things that is constant is the 'teaching a horse not to grab for food' which she describes in her first book.
Robin is grabbing at carrots. She walks up to him holding the carrot, if he lunges, she smartly turns her back and walks away. 5 secs later, she returns with the carrot. If he 'poses' but doesn't lunge, she C/R.
it seems to me that, forgetting the terminology of -R, +P or anything else, in the simple 'colder, hotter' of the game that is essentially what we're trying for - helping our horse to make choices that we want her to make because we, in our arrogance think this is best for them - 'turning and walking away'is a clear 'no' - as in, sorry, wrong choice. So if, in teaching basic polite behaviour, which is essential with any kind of food reinrocement, I pair my 'walking away' with a verbal 'I don't think so' (because i don't like 'no' either and this feels more friendly, and I want congruence of action, voice and feeling) then I have an effective signal that at least interrupts the behaviour of lunging for treats. WHen the horse learns to generalise, then 'I don't think so' ( or "I..." for short) acts as a useful interrupter when she thinks that my doing up the girth is a cue to move sideways (because I've reinforced a lot for light lateral movement in hand). She doesn't move, I can C/R for 'not moving' and we each begin to learn that there are subtleties to when a light pressure on the ribs is a request to move across.
Either way, it still seems to me that it's a 'no' and that if I only had a 'warmer' signal (either intermediate bridge or terminal bridge) then I'd be using half the tools in my tool box, and that a 'colder' signal is pretty useful to both of us.
BUT if Alex explained that I'm completely wrong, and given that we're all trying to learn what works best (where 'best' is defined as smoothly, effectively and with least stress for us both), I'm open to any thoughts.
Now, Catherine, in a new thread, would you like to talk to us about how to avoid frustration when using the clicker? seems a good topic for talk.
Sorry I haven’t replied earlier but I’ve been laid low by a horrendous cold and am just beginning to surface again…
I think Helen has already said some of this but here goes...
To be quite honest, although I mentioned the use of punishment, I can think of very few occasions when I would actually use it. I would never use punishment as an actual training tool, not least because it can’t be used to tell the subject what to do, only what not to do, and also because of the emotional and behavioural effects the use of punishment can create. However, there may be situations – for example if a horse is behaving aggressively and I don’t have enough information to know what is causing the aggression, or alternatively don’t have time to take any other action – where I might need to use punishment in order to avoid injury, either to myself or other people, or to prevent damage to property. As Murray Sidman points out in ‘Coercion and its fallout’, in such situations ‘…common sense tells us that we have to use whatever effective means are at hand.’ But as he goes on to say, the occasional emergency ‘may justify punishment as a treatment of last resort, but never as the treatment of choice. To use punishment occasionally as an act of desperation is not the same as advocating the use of punishment as a principle of behavior management.’
Punishment can undoubtedly be effective in dealing with certain forms of undesirable behaviour, but the problem is that in order to be effective it often needs to be far more severe than most people would feel comfortable about – and of course the timing has to be if anything even more precise than either +R or –R otherwise you risk creating not only confusion but massive resentment. I’ve seen horses (very often, though not always) stallions turn on their handlers because they have been unjustly punished once too often, and I’ve thought, ‘Good for you…’ but unfortunately the horse is then usually blamed for having a defective temperament; seldom does the handler question whether the horse’s behaviour might be the result of his or her own actions! Personally, on the rare occasions when I’ve actually used punishment I feel afterwards as if I’ve failed, even if the punishment actually worked, because I feel as if I should have been able to prevent whatever situation led to the use of punishment. And I invariably feel as if I’ve (temporarily at least) lost the trust of the animal concerned. So the only time I would use it is in the kind of situation cited above.
Sorry I can’t be more precise, but I felt I had to include punishment because if I said I never ever used it I would be fibbing! It just doesn’t form part of our normal horse management strategies. So I don’t know whether that’s answered your question at all…
Hi Leslie,
Thanks again for a really clear reply.
It seems to me that we're basically all on the same page re punishment. I concur with you completely Leslie that usually when you have to resort to it, it should be in an emergency type situation. I would also empathise with feelings of complete failure when you have to resort to it. However, I'm glad that most people here are saying that, although they do their very best not to use punishment, they won't rule it out completely. To me this is where a dose of common sense comes in that I've found lacking in so many clicker trainers, especially in the dog world.
Just one last thing about punishment - Helen, at the conference you talked about Monty 'punishing' the stand when he initiates join up by sending the horse away. Could you clarify whether you were being specific about Monty's really quite energetic sending away in that situation or do you mean that whenever you initiate a new behaviour, you always punish the previous one??? You see, this is where the semantics become REALLY important especially on e-mail. If you mean the latter, then I would have to hold my hand up and say that use a lot of +P. So would anyone who trains their horse since we would all have to 'punish' moving the horse from where he's stood to get our menage or whatever, before we could start our training. I think there's general confusion about this from reading some posts on the ABPC conference thread so it would be good if you could be really clear about this. To me, it's not punishing the current behaviour, its just initiating a new one but again I would have to say it all depends on what the horse thinks!!
Manda, really enjoyed your last post - Mulling over all this stuff makes me believe more and more that there are no absolutes in any of this, just horses with different personalities and thresholds for experienceing stress, pleasure etc. Whilst I think that a basic understanding of learning theory is good, I think that what really makes a good trainer is the ability to read a horse/dog and act appropriately. I like Alex's saying 'Go to people for opinions and horses for answers'
Thank you all,
Cath
I think that you are all correct and it is impressive how far the questioning and understanding of the science of behaviour has come in the last few years.
For me as a simple trainer I try to use the principles for Cannons Law. Simply put this means not using a more complicated explanation for behaviour is a simpler one exists.
For many years I have been saying it is not if we say no but rather how we say it that is important. I have come to believe that horsemanship and perhaps life is a balance of yes and no and that without no, yes has very little meaning or value. Without one we cannot judge the other.
If we ignore a behaviour we don’t want at that moment or withhold a click in hope of getting a better response we are saying no. When a behaviour doesn’t get the horses desired response we are saying no that didn’t work. I as many of you know, use the training game to demonstrate learning and during that game many of you have “asked a question”, does this work by repeating a behaviour such as picking something up. When there was no click you knew I was saying no that’s not it.
We can and indeed must be ready when required to say no to the domesticated horse for safety. However, for me all we have to do is listen to the horses they are telling us what works and what doesn’t, they tell us how they feel about no. I as you know use as much +R as possible, but the truth I accept is that as hard as I try the horse will still perceive some of my behaviour as negative, mainly because I am often changing what works for them.
The tools of no are consistency, persistence and patience but above all lightness. Often I meet owners who worry about no because they themselves feel over controlled in their lives, they feel they are told no too much themselves or they punish their own mistakes with a mental no and this leads them to wanting to give their horse total freedom and only yes so they are loved and are not resented.
No is feed back it occurs in nature and in domestication and as always it is a matter of perception, for one a slap on the nose, whip on the bum is an acceptable no and for others evening saying no is too harsh, why is that?
Horses use both yes and no, but we are not horses and easily make mistakes in communication just as we do with other humans. So the consistency of our individual no is what matters.
And just as I am here with reference to me throwing the cone to far for the horse to reach. My mistake, an error in judgement, rather than use a continuous stream of unnatural sound to make the horse do what I wanted I would preferred to let the horse think, choose and decide and if required I should have reshaped the behaviour to allow the horse to make decisions by moving the cone closer. For me it is not about making the horse touch the cone its about allowing it to learn it can. We all talk a great deal about what’s natural but horses never xxxxxxx each other as far as I am aware. I have to say though, that now I have developed different CT methods and the use of intermediate and terminal bridges I could have help the horse more if required but for me its about the animal learning and developing problem solving and if I direct the animal constantly and solve all its problems it does have to learn to think.
Listen to your horse as an individual it will tell you if you are ok or not, but never try to be a horse. It is not our intention that is important but the horses perception or our action that determines whether it is +R –R or P we have just used.
Why do some horses get frustrated with clicker training? Perhaps because there are so many silent no’s which the horse doesn’t like as it wants yes. Oh my god may be clicker training is abusive
Cath said: “You see, whenever people talk about using 'no', its sometimes hard not to sound as if you beat the horse with a big stick!!! I think there are ways of using 'no' that don't in any way equate to this and don't cause stress. I actually thought that the training game you demonstrated possibly highlighted this point although I don't think it's what you intended. It was unfortunate that the lady who volunteered to be the trainer had such a nice, quiet 'no'. You see to me, the level of 'no' that she used looked a bit like that game you play as a kid where you hide something and then say 'warmer', 'hot' 'cold' to the person doing the finding. To me, this is what her 'no' was like. I actually think there is an equivalent with animals and I think it helps to make things clear to them. It's certainly the attitude I try to have when I train. The horse/dog has a puzzle to solve and I will help by saying hot, warmer and sometimes 'cold'.
There is also the 'no' you talked about above used in rare circumstances as a punisher. I totally agree that these should be rare instances and dependent on the horse and the situation.”
Cath, ‘no’ by definition is something that lessens a behaviour, therefore it is a punisher, whether positive or negative depends on how it is introduced. This can be by the removal of food or attention, or by the addition of something unpleasant that causes discomfort. By definition it is aversive to the animal, it is something they wish to avoid. However you are totally right to say there is a continuum of aversives. The weaker the aversive then I would imagine the less stressful it is for the animal. But still something that they wish to avoid, or else it wouldn’t actually be effective. Therefore there will always be an element of stress involved. This is why I do not choose to use ‘no’ when I am training a behaviour. And as Lesley rightly points out there is fall out when using punishment, the most noticeable of which is avoidance. The training game as we did it at the conference was just a snippet and not long enough to do it justice or to really go into the fine details, which was a shame, obviously when we do this again we will have to make it two days not one! It doesn’t matter that we were working with people, the whole point was to get people thinking about the impact they are having on their animals. Remember what the person being trained said about how they felt about the different methods and then imagine your horse in that position. We demonstrated positive punishment and positive reinforcement. The next step would have been to look at negative reinforcment and at negative punishment. That might perhaps have made things clearer to everyone. Particularly considering the context in which we did it, when during the mornings lecture we covered stress and the various behavioural modification tools for dealing with problem behaviour. We did discuss the lady’s attitude to saying ‘no’ and how with different individuals that would have been different, and the effect this then might have had on the individual being trained. I also recommended that everyone went away and played the training game with as many different groups of people as they could, and got feedback from them, so as to begin to better understand what it is like both to train and be trained.
Cath also says: ” You see, if you think about it, 'no', 'yes' whatever, only counts if the animal understands your intent etc. You can do all the +R in the world but if the animal feels stressed, intimidated, dominated, it's no better than using -R or +P.”
Well this is the point we are trying to get across. +R is not a training method! It is just the term used for rewarding behaviour. And it is only rewarding if the animal views it to be so, not just because the trainer intends it to be. Therefore a reward should not create stress, intimidation etc. Bad training creates stress, intimidation etc. Certainly many trainers who claim to use +R can cause these things. But it is not because they are using +R, it is because of HOW they are using it. In exactly the same way as there are trainers out there who use –R who are actually good trainers, who stress and intimidate their horses very little.
How many times do I have to say it? I feel like a broken record! We should all be striving to be GOOD trainers. And GOOD trainers understand how to use +/-R/P and the impact it has on the animal. GOOD trainers stress the animal as little as possible if at all. GOOD trainers are consistent, predictable and at all times understand what is going on with the animal they are training. I would never call myself a ‘positive reinforcement trainer’ because this would be a lie. In the same way I don’t advertise myself as a ‘clicker trainer’, rather I am someone who uses the clicker extensively when training. However I would hope that I am doing my best to be as GOOD a trainer as I can manage with the knowledge and experience I have so far in my life- and that I will ALWAYS be striving to be a BETTER trainer!!!
Sorry Cath this is not a rant against you AT ALL, just an opportunity for me to jump on my soap box AGAIN!!! Just a good opportunity to use your quotes to make a point.
Manda I don’t think I need to do an individual reply to you cause I think I said it all above!!
Cath said “Just one last thing about punishment - Helen, at the conference you talked about Monty 'punishing' the stand when he initiates join up by sending the horse away. Could you clarify whether you were being specific about Monty's really quite energetic sending away in that situation or do you mean that whenever you initiate a new behaviour, you always punish the previous one???”
Good question Cath. AS Catherine (B) rightly pointed out at the conference, -/+ R/P apply to the effects they have on the behaviour, not the actions themselves. So for example with a dog that jumps up excessively, yet people push him down and he still keeps jumping up- they INTEND the pushing down to be a punisher, because they want to stop him, however he finds any physical contact a pleasure and wants attention, so he views the pushing as a REWARD. We know this because he keeps on repeating the behaviour- in other words it has increased, rather than decreased. I was being specific about Monty’s strong send away when I said he was punishing the stand- he applied an aversive (lots of movement, which causes the horse to feel fear instinctively) and the horse responded with a natural fear response which is flight. However to know whether this aversive is effective as a punisher we would have to look at the horse in the context of the start of a second, third and fourth join up- as soon as the rope is unclipped, does the horse immediately and easily move away, or does Monty have to use more and more pressure to send him away? I have to write more about this on the APBC thread in reply to Sarah so I won’t do it here! But I hope this helps you a bit. I suppose with respect to your question do you always punish the previous behaviour when you initiate a new one (I take it you mean by using negative reinforcement) well, both negative reinforcement and positive punishment involve the animal experiencing an aversive. So it depends on the effect they are having on the behaviour whether or not they actually are punishing or not, but at the end of the day the animal still experiences an aversive and therefore a degree of stress depending on the strength of the aversive. It is what the animal is experiencing that you should be thinking of, and remembering everything we talked about at the conference about stress! That is why trainers should think carefully when training using solely negative reinforcement! I’m not saying don’t use it, but be AWARE of how it works!
Ben says “For me as a simple trainer I try to use the principles for Cannons Law. Simply put this means not using a more complicated explanation for behaviour is a simpler one exists.”
Thanks Ben for this, we did talk about the principle of parsimony on the day, I totally agree with you.
”For many years I have been saying it is not if we say no but rather how we say it that is important.”
Excellent point and what I was trying to say above!!
Ben thank you so much for your post, lovely to have some clear thinking and common sense! If only more horse trainers thought like you.
“but for me its about the animal learning and developing problem solving and if I direct the animal constantly and solve all its problems it does not have to learn to think.”
I think this is a very good point- it is something I talk about a lot about at my workshops (probably thanks to both you and Heather Simpson getting me thinking about it!)- and I think Karen Pryor makes the point really well in her book Lads Before the Wind where she talks about how most horse training does not encourage horses to think, in fact it does the very opposite, it encourages them to act without thinking. This is what I love about being able to use a tool like the clicker, and training using reward, that beautiful moment when you see a horse first begin to open up and try something new, to think about the consequences of their actions and use those thoughts to develop new behaviours, with enthusiasm and enjoyment.
”Listen to your horse as an individual it will tell you if you are ok or not, but never try to be a horse. It is not our intention that is important but the horses perception or our action that determines whether it is +R –R or P we have just used.”
You know I agree with this, I think I already said it up there somewhere!
”Why do some horses get frustrated with clicker training? Perhaps because there are so many silent no’s which the horse doesn’t like as it wants yes. Oh my god may be clicker training is abusive ”
Oh no!!
Helen,
First of all, thankyou for clarifying what you meant about punishment in the context of Monty's stuff and I look forward to reading more about it on the other thread. Second, I didn't take anything you said as a rant so don't worry.
Helen said:
Cath, ‘no’ by definition is something that lessens a behaviour, therefore it is a punisher, whether positive or negative depends on how it is introduced. This can be by the removal of food or attention, or by the addition of something unpleasant that causes discomfort. By definition it is aversive to the animal, it is something they wish to avoid. However you are totally right to say there is a continuum of aversives. The weaker the aversive then I would imagine the less stressful it is for the animal. But still something that they wish to avoid, or else it wouldn’t actually be effective. Therefore there will always be an element of stress involved. This is why I do not choose to use ‘no’ when I am training a behaviour.
Believe it or not Helen, I do actually understand this. But if you re-read my last post, and Bens' post I think you'll agree that there is concensus that even not clicking can be construed as a 'no';. This is what I meant when I talked in terms of 'walmer', 'hot' etc
Helen then said:
Remember what the person being trained said about how they felt about the different methods and then imagine your horse in that position. We demonstrated positive punishment and positive reinforcement.
Unfortunately, the person being trained was Jenni who undoubtedly knew the point you were trying to make. (I'm NOT saying it was a deliberate set up - just unfortunate that nobody else volunteered). Jenni said that she didn't feel as confident when the trainer was using 'no'. However, I would ask you to consider whether kids look stressed by the walmer, hot etc game. (OK, I know kids don't really play this game anymore - they have gameboys but you know what I mean). I remember playing this as a kid and certainly don't remember feeling anything but pleasure - DEPENDING on who I was playing with!
Ben said:
For many years I have been saying it is not if we say no but rather how we say it that is important.”
This is EXACTLY my point. Thank you Ben.
I've played the training game in the past many times with friends who either own dogs, horses or no animals at all. I deliberately practice this to get my own attitude right. I have never once had someone say they felt stressed by it. I believe my animals are also of the same opinion, though of course I can't ask them.
Everybody in their own way seems to be saying that it's not what you do, its how the animal regards it. Couldn't agree more. Helen, you talk a lot about stress and how we should pick methods that don't stress animals. Also, couldn't agree more. However, I go back to something I said about million posts ago, in my opinion, what's clearest is probably least stressful long term.
respectfully,
Cath
Well I can't remember if I said specifically 'pick methods that don't stress animals' and if I did then I am sorry because I should have said 'pick methods that don't stress animals any more than necessary'. I live in the real world and stress happens!
Am surprised you have never had anyone say they feel stressed during the training game- I often find people saying that and also have those that are playing the part of 'trainers' saying they are stressed according to what method they are using, notably one person who just couldn't bring herself to say 'no' and another person who found it very frustrating to work out how to reward a behaviour.
However aside from those points I think we are going round and round in circles here- surely we are all now agreed about what punishment is, the effect it has on the animals, and when and when not to use it!
And since the original debate was about what's natural for a horse and whether they ever use reward, well, we know they indulge in behaviours that they find reinforcing because they find them pleasurable, so that is certainly not a reason to only train them using negative reinforcement and punishment! So surely now we are all agreed that we wish to train in order to minimise stress for the horse, it seems likely that it is best to train using as much positive reinforcement as possible, with punishment only where absolutely necessary, and occassional mild negative reinforcement.
(And now just to throw the cat among the pigeons, I don't believe that witholding rewards when you are shaping behaviour ought to be that stressful if your shaping plan is good enough!! Obviously there is a degree of stress involved, but it should be possible to keep that to a minimum).
Helen
