Setting up an OC behavior

[Hawkingcolorado]

For you guys who are using OC. What are some examples of a behavior you wanted, or didn't want, and how did you set it up? Is this even the correct term? I don't know that much about OC yet, but am interested. Ive got "Don't shoot the Dog" on the way from Amazon...for starters...

[Keith Denman]

Get Ken Blanchards book Whale Done. It explains it in management terms but applies the same. It can be read in half an hour and you can apply it immediately. OC is rewarding the good immediately and not giving the bad habits rewards. Also watch the dog whisper he uses it. Get the book Whale Done it is the best book I have read on OC.

[kitana]

There are many ways to train a behavior using operant conditionning, mainly capturing, free-shaping, luring and targeting. You already use OC without knowing it, and classical conditionning as well.

Typically in OC, the animal does a behavior that is reinforced by a reward, thus being called positive (you add a consequence, positive in the mathematical sense) reinforcement (to make the behavior occur more frequently) or R+. There are other quadrants to this but they aren't used as much as positive reinforcement.

To be more efficient, you can pair a signal with the R+ to tell the bird precisely which behavior is reinforced. By doing so you eliminate the need to give the R+ at the exact moment the hawk does the behavior, you gain about 3-5 seconds to deliver the reward. The clicker and whistles are the traditionnal signals, but a Yes or cluck of the tongue can be used.

So back to training a behavior. The first way to train a behavior is called capturing: you capture the whole, finished behavior as it is. You click/treat when everything is perfect. It can be used for very fast behaviors such as eye wink, tongue flick, etc. The conditions to use capturing is that the behavior needs to naturally happen very often. It's almost a waste of time to use capturing with complex behaviors that the hawk won't do spontaneously often enough for you to reinforce it.

For complex behaviors there is shaping, or free shaping. You will click/treat a tiny step that will lead to the desired, finished behavior. Example; you want your hawk to look away from you, over its left shoulder, for 5 seconds. Good luck if you try to capture that!! But by reinforcing the hawk for looking a little bit on the left, and then waiting for him to look a little bit further, and further, and then maintain the position for half a second, one second, etc, you'll obtain the desired behavior.

One can also use luring. Luring means using food as a lure (not in the falconry sense) to induce a behavior. It is very commonly used in falconry: place food on glove and the hawk goes upon glove, place food on perch and the hawk goes on the perch, etc. However luring can become a problem if the animal learns to do the behavior only when seeing the lure... Typically trainers will try never to lure more than 3 times, they will fade the lure very quickly, but it can quickly become a problem.

Then there is targeting. With targeting, the animal learns to touch or follow an object called a target. It can be a laser pointer also. then you use the target to teach new behaviors, and you fade your target so it is not needed with the final behavior. As an example, one could use a laser pointer or a target glued in the back of the giant hood to teach the hawk to enter the giant hood. Or teach a hawk to hood itself by targeting the hood and using free shaping to shape a "enter your head in the target" behavior.

My answer was probably too long or too in depth for what you wanted, but that's a recapitulation of the basis of OC training. I would not suggest to read Don't shoot the dog or Ken Blanchard if you wanted to start training that way: they are excellent books, but more focused on the philosophy/inner workings of OC. I'd go directly to a practical book like any Karen Pryor/Morgan Spector books on clicker training, even one for dogs. The principles are illustrated in exercises that can easily be modified to fit a hawk.

Oh yeah: it's primordial to have a very high rate of reinforcement of you don't want the animal to become frustrated and either give up or even attack out of frustration! You have to choose your criterions in a way that garantees success to the animal, and work your way from that in tiny increments. It is what gives most people difficulties, that and the timing of the marker signal.

[Hawkingcolorado]

Kitana,
Thanks very much! That was pretty much exactly what I was wondering about..To Have to wait for the animal to do something it might do only once a week, then reward it, seems impossible! So more natural and common "good' behaviors are easier to do OC with. Such as feeding a young Gos when it's relaxed, as Paul mentioned(I think)....Some thing like flying out in a circle and returing to the glove evolves fron RP's? Is that right?.....Thanks

[Keith Denman]

Hi Audrey:
That is a very good explanation. I never heard of the other writers you mentioned I will have to check them out.

[Mandragen]

Ive got "Don't shoot the Dog" on the way from Amazon...for starters...

Excellent book, should be a required read for all trainers.

I haven't heard of the other one yet, I'll have to get a hold of a copy!

[Mandragen]

So back to training a behavior. The first way to train a behavior is called capturing: you capture the whole, finished behavior as it is. You click/treat when everything is perfect. It can be used for very fast behaviors such as eye wink, tongue flick, etc. The conditions to use capturing is that the behavior needs to naturally happen very often. It's almost a waste of time to use capturing with complex behaviors that the hawk won't do spontaneously often enough for you to reinforce it.


For example, when my dog has a bug flying around his head, he smacks at it with his mouth. So we taught him to "blow kisses" by reinforcing that behavior one day. Just use your bridge(whatever your noise is to bridge that gap), with my dog it's "good boy!", and then give that reinforcment. At first they look like, "What was that for?" and then they start to catch on. Once it got the picture and started smacking away, just introduce the discriminative stimulus you plan to use, in this case it was "Blow kisses!" (with a kiss action blown to him), and then when he does it, bridge and reinforce.

[Mandragen]

Some thing like flying out in a circle and returing to the glove evolves fron RP's? Is that right?.....Thanks


If I were gonna go for a desired behavior like that, I would teach a target. Usually the easiest target to make is a length of PVC and stick a round object at one end. Make sure you select the proper size to fit the animal, an elephant target pole is going to be much larger than a bird target pole.

To simplify:
Teach target by touching target to beak, bridge and reinforce. Once it understands wait until bird touches target, bridge reinforce. Then, just like flying to glove make it fly to target, bridge reinforce. Then you get a couple targets and people to help, having people present them in the desired pattern. This is exactly how we get dolphins to swim in patterns. Of course it can take some shaping in between and there is a bit more to it, but that gives you the idea.

[goshawkr]

Somehow I knew you were gunna make your keyboard sore when I saw this thread pop up ...... :D


There are many ways to train a behavior using operant conditionning, mainly capturing, free-shaping, luring and targeting. You already use OC without knowing it, and classical conditionning as well.

Typically in OC, the animal does a behavior that is reinforced by a reward, thus being called positive (you add a consequence, positive in the mathematical sense) reinforcement (to make the behavior occur more frequently) or R+. There are other quadrants to this but they aren't used as much as positive reinforcement.

To be more efficient, you can pair a signal with the R+ to tell the bird precisely which behavior is reinforced. By doing so you eliminate the need to give the R+ at the exact moment the hawk does the behavior, you gain about 3-5 seconds to deliver the reward. The clicker and whistles are the traditionnal signals, but a Yes or cluck of the tongue can be used.

So back to training a behavior. The first way to train a behavior is called capturing: you capture the whole, finished behavior as it is. You click/treat when everything is perfect. It can be used for very fast behaviors such as eye wink, tongue flick, etc. The conditions to use capturing is that the behavior needs to naturally happen very often. It's almost a waste of time to use capturing with complex behaviors that the hawk won't do spontaneously often enough for you to reinforce it.

For complex behaviors there is shaping, or free shaping. You will click/treat a tiny step that will lead to the desired, finished behavior. Example; you want your hawk to look away from you, over its left shoulder, for 5 seconds. Good luck if you try to capture that!! But by reinforcing the hawk for looking a little bit on the left, and then waiting for him to look a little bit further, and further, and then maintain the position for half a second, one second, etc, you'll obtain the desired behavior.

One can also use luring. Luring means using food as a lure (not in the falconry sense) to induce a behavior. It is very commonly used in falconry: place food on glove and the hawk goes upon glove, place food on perch and the hawk goes on the perch, etc. However luring can become a problem if the animal learns to do the behavior only when seeing the lure... Typically trainers will try never to lure more than 3 times, they will fade the lure very quickly, but it can quickly become a problem.

Then there is targeting. With targeting, the animal learns to touch or follow an object called a target. It can be a laser pointer also. then you use the target to teach new behaviors, and you fade your target so it is not needed with the final behavior. As an example, one could use a laser pointer or a target glued in the back of the giant hood to teach the hawk to enter the giant hood. Or teach a hawk to hood itself by targeting the hood and using free shaping to shape a "enter your head in the target" behavior.

My answer was probably too long or too in depth for what you wanted, but that's a recapitulation of the basis of OC training. I would not suggest to read Don't shoot the dog or Ken Blanchard if you wanted to start training that way: they are excellent books, but more focused on the philosophy/inner workings of OC. I'd go directly to a practical book like any Karen Pryor/Morgan Spector books on clicker training, even one for dogs. The principles are illustrated in exercises that can easily be modified to fit a hawk.

Oh yeah: it's primordial to have a very high rate of reinforcement of you don't want the animal to become frustrated and either give up or even attack out of frustration! You have to choose your criterions in a way that garantees success to the animal, and work your way from that in tiny increments. It is what gives most people difficulties, that and the timing of the marker signal.

[kitana]

Geoff, you should be praising me for my self-control instead: did you see when the thread was started, and when I posted??? I had to refrain posting all night long until I got time this morning... THAT is a feat! ;)

[goshawkr]

Geoff, you should be praising me for my self-control instead: did you see when the thread was started, and when I posted??? I had to refrain posting all night long until I got time this morning... THAT is a feat! ;)

Indeed I should be praising you for that. I actually hadnt noticed the restaint you showed, but now that you point it out I do wonder, how ever did you sleep?? :D

And the thread isnt closed yet - I think I should pop an asprin in the mail for you to give to your keyboard. :D

BTW - I neglected to mention that your post was excellent, as usual.

[Hawkingcolorado]

I want more info!!!:D

[kitana]

I want more info!!!:D

Oh boy... Only if you pay me a new keyboard... And aspirins... ;)

[goshawkr]

Oh boy... Only if you pay me a new keyboard... And aspirins... ;)

If only we had a seperate forum for OC falconry here on NAFEX maybe you could post that info you wrote up on how you trained your harris' hawk.

[MourneMan]

Very informative please keep it going.
Malcolm

[goshawkr]

I want more info!!!:D

OC training can encompass some very advanced concepts. However, like all other things you can and should start simple. Boiled down to its most basic, OC is a way that you can "mark" a behevior with a signal (Conditioned Reinforcer or CR) that in effect means "Yes! that is exactly what I want!!" which can be quickly delivered. Any signal can work, and the foundation before you do any training is to make the link between the CR and the reward, by giving the CR and immediately giving a reward several times. You know the link is made when the body language clearly indicates they are looking for the reward as soon as the CR is delivered.

The CR lets you decouple the delivery of a reward with a behavior you wanted. With practice, this will let you reach out and "grab" behaviors that are fleeting - even as short as a second. In practicle application, this also means that if your skill is off, you end up rewarding something other than what you intended because you missed the timing and "marked" something other than your goal.

One thing that is very different from OC training as opposed to most other training methods (even positive ones) is that you are not coaxing, nudging or anything else. At least not usually - you still can nudge or coax and the use OC. But most of the time, you sit patiently and wait for the behavior you want to see to occur on its own, then reward it.

Once its established and well entrenchend, then you add a cue to put it on command.

This is upside down from the "normal" approach.

For example, a common "recipie" to teach a dog to sit is to get the dog in front of you, force its butt to the ground as you say "sit" and then give it praise and a treat. Repeat until the link is made between the word and the act. This works.... but OC handles it different. To teach a dog to sit using OC, you just hang out near the dog until it sits and then give the CR as soon as its butt hits the ground and give it a treat when you can (up to 15 minutes later). Spend a few 15 minute sessions doing this, and THEN make an association to the "sit" command by only giving the CR if you have asked the dog to sit before hand.

I intended this as a pre-amble for a real falconry solution I used OC for that I had posted to another forum, but my intro got a bit wordy. I'll do another post with that example.

The simplest way to wrap your head around OC is to view it as a collection of training tools. Sometimes some of those tools work very well for a given situation. Sometimes more traditional approaches will be faster and/or more effective. Sometimes when you come up with a solution you'll use just one OC related tool, and sometimes you'll need more than one.

[goshawkr]

Ok, here is a simple "how it happened" that I had previously posted on another forum.

I had a goshawk that didn't take to going into the giant hood very well. Each time I went hunting, the argument about going into the giant hood got more pronounced until I was having to cast him, toss him and slam the door shut before he could get out. Yes, with a goshawk. What always surprised me is that he was always calm and ready to go when we got to the field - I expected a basket case and to have to turn around and go home afterwards. Having a goshawk argue about going into the giant hood wasn't a surprise to me, but getting progressively and consistently worse was. I decided to turn things around on the day I had to cast him to get him in, and came up with an OC recipe to do it.

Its been a while since this happened, so I forget exactly what occurred during which training session. Fundamentally that dosnt matter. I do know this guy was amazing in how fast and how well he would play along with OC games. A lot of goshawks like to be stubborn and manipulate back, but not him. In two 15 minute sessions he was willing going inside the giant hood. On the third session, he would just sit calmly inside waiting for a tidbit.


To begin with, he was perched in a room on a tail saver perch with lots of leash. I placed the Giant hood on the floor with the door open close enough that he could get inside, and I sat down with some tidbits handy. Here are the progression of goals we did. Each time he met a goal, I would give the Conditioned Reinforcer to mark that instant in time and toss a tidbit on the ground for him. As he got good and solid with a goal, we moved to the next goal on the list.
Look in the general direction of the giant hood while sitting on the perch.
Look directly at the giant hood while sitting on the perch.
Move in the general direction of the giant hood.
Move directly towards the giant hood.
Move closer to the giant hood than you have been doing before.
Put your head inside the giant hood.
Put your body in the giant hood.
Perch inside the giant hood.
Hang out inside the giant hood longer than you have been doing before.
One key thing to keep in mind is to not get bogged down on the "step" you are on, and understand you are working towards the ultimate goal. If progress is made towards a goal further up the list than where you are, REWARD IT! Reward it more than normal. If the critter you are training is ready to skip ahead that is a good thing. On the other hand, if progress is going very slowly or not at all, then maybe you need to step back to a previous goal, at least for a few minutes if not for a whole session. The rewards should come at least once every 2-3 minutes.

There are of course times when the steps are rigid, but that would be defined by the ultimate goal you are training for.

Its also important to reinforce the ultimate big goal with a lot of time/repetition. In my example above, we got to step 9 on the second day, but I still set up 4-5 training sessions that just consisted of my goshawk perching in the giant hood with the door open (while tied to the tail saver perch) and me wandering through the room every once in a while to toss him a tidbit if he was sitting in the giant hood calmly. He could sit anywhere he wanted on those sessions, but the only thing he could do to con me out of a tidbit was to sit inside the giant hood.

One funny thing that happened during this, and it was clear insight that he knew the game, is that while he was on step 4, after his second tidbit at that goal level he marched directly towards the giant hood, looked over his shoulder DIRECTLY at me, and put his foot in it.

There are plenty of ways to condition a hawk to using the giant hood - and they all work.

The beauty of OC however, is that it can be used to teach anything. Its like some magical "do anything" tool. The only limits are the imagination of the trainer to come up with the goal list, and the physical and mental capacity of the animal your teaching. If the animal is capable of performing the act, you can teach it using OC. In the case above, I was trying to teach the goshawk to accept the giant hood by general conditioning, but when that was ineffective, I resorted to OC to turn things around quickly and effectively.

Actually, when Steve Layman was in college learning OC as a psychology major he had the great insight to realize that the Conditioned Reinforcer is a simple one word language. It means "Yes!" or "Do that!" And to take this even further and more complicated, you can also fold in another concept from a competing psychological theory (Aversion training) and incorporate something akin to "No!" or "Don't do that!", although this is a bit delicate to pull off with an animal like a hawk.

With an animal actively trying to get you to give them the reward, you can accomplish a LOT with just that one (or two) word language. It can be like having a remote control wired to their brain.

[Hawkingcolorado]

Ok, here is a simple "how it happened" that I had previously posted on another forum.

I had a goshawk that didn't take to going into the giant hood very well. Each time I went hunting, the argument about going into the giant hood got more pronounced until I was having to cast him, toss him and slam the door shut before he could get out. Yes, with a goshawk. What always surprised me is that he was always calm and ready to go when we got to the field - I expected a basket case and to have to turn around and go home afterwards. Having a goshawk argue about going into the giant hood wasn't a surprise to me, but getting progressively and consistently worse was. I decided to turn things around on the day I had to cast him to get him in, and came up with an OC recipe to do it.

Its been a while since this happened, so I forget exactly what occurred during which training session. Fundamentally that dosnt matter. I do know this guy was amazing in how fast and how well he would play along with OC games. A lot of goshawks like to be stubborn and manipulate back, but not him. In two 15 minute sessions he was willing going inside the giant hood. On the third session, he would just sit calmly inside waiting for a tidbit.


To begin with, he was perched in a room on a tail saver perch with lots of leash. I placed the Giant hood on the floor with the door open close enough that he could get inside, and I sat down with some tidbits handy. Here are the progression of goals we did. Each time he met a goal, I would give the Conditioned Reinforcer to mark that instant in time and toss a tidbit on the ground for him. As he got good and solid with a goal, we moved to the next goal on the list.

Look in the general direction of the giant hood while sitting on the perch.
Look directly at the giant hood while sitting on the perch.
Move in the general direction of the giant hood.
Move directly towards the giant hood.
Move closer to the giant hood than you have been doing before.
Put your head inside the giant hood.
Put your body in the giant hood.
Perch inside the giant hood.
Hang out inside the giant hood longer than you have been doing before.

One key thing to keep in mind is to not get bogged down on the "step" you are on, and understand you are working towards the ultimate goal. If progress is made towards a goal further up the list than where you are, REWARD IT! Reward it more than normal. If the critter you are training is ready to skip ahead that is a good thing. On the other hand, if progress is going very slowly or not at all, then maybe you need to step back to a previous goal, at least for a few minutes if not for a whole session. The rewards should come at least once every 2-3 minutes.

There are of course times when the steps are rigid, but that would be defined by the ultimate goal you are training for.

Its also important to reinforce the ultimate big goal with a lot of time/repetition. In my example above, we got to step 9 on the second day, but I still set up 4-5 training sessions that just consisted of my goshawk perching in the giant hood with the door open (while tied to the tail saver perch) and me wandering through the room every once in a while to toss him a tidbit if he was sitting in the giant hood calmly. He could sit anywhere he wanted on those sessions, but the only thing he could do to con me out of a tidbit was to sit inside the giant hood.

One funny thing that happened during this, and it was clear insight that he knew the game, is that while he was on step 4, after his second tidbit at that goal level he marched directly towards the giant hood, looked over his shoulder DIRECTLY at me, and put his foot in it.

There are plenty of ways to condition a hawk to using the giant hood - and they all work.

The beauty of OC however, is that it can be used to teach anything. Its like some magical "do anything" tool. The only limits are the imagination of the trainer to come up with the goal list, and the physical and mental capacity of the animal your teaching. If the animal is capable of performing the act, you can teach it using OC. In the case above, I was trying to teach the goshawk to accept the giant hood by general conditioning, but when that was ineffective, I resorted to OC to turn things around quickly and effectively.

Actually, when Steve Layman was in college learning OC as a psychology major he had the great insight to realize that the Conditioned Reinforcer is a simple one word language. It means "Yes!" or "Do that!" And to take this even further and more complicated, you can also fold in another concept from a competing psychological theory (Aversion training) and incorporate something akin to "No!" or "Don't do that!", although this is a bit delicate to pull off with an animal like a hawk.

With an animal actively trying to get you to give them the reward, you can accomplish a LOT with just that one (or two) word language. It can be like having a remote control wired to their brain.

Great example!! Thanks very much...

[kitana]

Really well explained, Geoff. Good example with the giant hood: I need to do some more work on that with my HH and I never thought about tethering him during the session, I worked him attached to the glove and I delivered the tidbits from a hole I made in the GH, but it would work waaaay better if tethered.

One of my techs is taking the Karen Pryor clicker course, she'll be the first one in Quebec to make it. She will have her final exam next week, and her skills increased immensely already. One thing she learned that have changed from when I first learned OC 5-6 years ago, is that now the emphasis is put on putting the behavior on cue as fast as possible, and then refining it. She was taught to put a totally new behavior on cue in 20 tidbits! Crazy stuff really. If someday I have 5000$ that I don't know what to do with, I'll take that course for sure. I don't feel like I need it for my future accipiter, but it wouldn't hurt!!

[kitana]

Oh and another detail, that is important when working with hawks: Jesus Rozalez-Ruiz demonstrated that even with clicker-savvy animals, the time between the CR and the reward should never exceed 5 seconds. When he did exceed 5 seconds, he saw unwanted behaviors crop up, especially frustration and it's twin, aggression. So it's only us humans who will work our a""es off for 2 weeks of time before receiving our reward... No other animal is as dumb as us... lol

[Mandragen]

If we are diving into this wonderful world of OC, one thing to mention at this point are superstitious behaviors. Make sure that you are only reinforcing the exact behavior that you want! It is possible for your subject to be doing the desired behavior but also exhibiting other undesired behaviors at the same time.

If we take the example above (which is excellent by the way, nice write up!), when that hawk is in that box for the first time reward that. But as stated it should be kept up with multiple session, and if you want a really strong behavior, even when it is a known behavior it should be done randomly all the time. During the sessions pay attention to the little behaviors as well, not just "the hawk is in the box". The hawk needs to be sitting nice, calm, quiet, and still. There are multiple criteria for a behavior to be correct.

We used to have 5+ sea lions that would line up at a fence, which allowed one trainer to feed 5+ sea lions at one time. One of them who's position was at the end and sort of in one's peripheral vision, would sometime get overlooked if the trainer wasn't paying attention. The lion learned real quick that if it moved it got the trainers attention. Now it has a huge problem with swaying back and forth the whole time it sits there, kind of obnoxious really.

[goshawkr]

Really well explained, Geoff. Good example with the giant hood: I need to do some more work on that with my HH and I never thought about tethering him during the session, I worked him attached to the glove and I delivered the tidbits from a hole I made in the GH, but it would work waaaay better if tethered.

Have you tried the "tradional" approach of tossing a large chunck of meat (rabbit leg, quail) into the giant hood and letting your hawk eat in it? Its minimal work and generally works very well, especially during the "off" season when your not actively hunting.

It certainly isnt OC, but it gets thejob done. 80)


One thing she learned that have changed from when I first learned OC 5-6 years ago, is that now the emphasis is put on putting the behavior on cue as fast as possible, and then refining it. She was taught to put a totally new behavior on cue in 20 tidbits! Crazy stuff really. If someday I have 5000$ that I don't know what to do with, I'll take that course for sure. I don't feel like I need it for my future accipiter, but it wouldn't hurt!!

Interesting.

Very few of the items that I train my hawks to do involves a cue, but perhaps I should put it on cue anyway.

I have a project to train some delightfully irreverant behaviors to a goshawk sometime.

[Hawkingcolorado]

After using a tidbit sitting in the box to get my HH to go in with no stress,
I tought I'd try and gte her to come back to the truck if we got skunked hunting.
I opened the tail gate flipped up the topper, and openen the hawkbox door. She came over in about 2 minutes when she saw me just sitting there. Now she does it right away when she knows we're done hunting...

[goshawkr]

Oh and another detail, that is important when working with hawks: Jesus Rozalez-Ruiz demonstrated that even with clicker-savvy animals, the time between the CR and the reward should never exceed 5 seconds. When he did exceed 5 seconds, he saw unwanted behaviors crop up, especially frustration and it's twin, aggression. So it's only us humans who will work our a""es off for 2 weeks of time before receiving our reward... No other animal is as dumb as us... lol

Well, the base theory is that there is a 15 minute window between when the CR is given and when the reward has to be delivered for the connection to made. Thats rooted on the underlying research by Skinner.

You are correct to point out that there is an extra difficulty with a keyed up hawk. Especially one at flying weight. However, even so, there are ways to work through this.

The simplest is to find ways other than handing the tidbits. They can get keyed up all they want, and go get a tidbit tossed on the ground for example, and nobody gets hurt and painful behaviors are not rewarded.

I do this frequently.

Another I commonly do is to get cocky about my quick reflexes, and place teh tidbit on the glove and get my hand out of the way before the hawk can foot me. Its not really as hard as it sounds, but it does take very keen attention to what the focus of the hawk is. They are quicker than we are, and if its keyed on your hand before you make the move you will never ever make it no matter how fast you are. particulaly with an accipiter.

However, the one I really prefer is to shape manners. The hawk has to present the behavior to be rewarded - then I deliver the CR. It then must maintain a polite bearing as I deliver the reward, or the reward goes away. I try to get this reinforced by withdrawing the tidbit at the first sign of aggression, then trying to present it a few moments later after the hawk acts calm. It is a very delicate dance and frequently when I am working with a new hawk I get a new scar or two. I never claim to be particularly bright. :D But they do come around to this (sometimes not before I decided to avoid the pain). Its pretty simple for them - calm = appearing tidbit and un-calm=retreating tidbit.

I know Steve Layman's name regularly comes up with OC and falconry, and not without justificiation (he was one of the first to apply OC to falconry). A real illustration on how far this last method I mentioned can progress was when I watched him playing some shaping games with his goshawk to get it in the "mental mode" for hunting. He would give the CR and then without restraining the jesses at all hold out his bare hand with a rabbit leg in it so the hawk could tear off a tidbit.

I have a goshawk that will gently take a tidbit from my bare hand, then lick the blood off my finger. But even she wont do what Layman's hawk did without hurting someone.

[Mandragen]

Interesting.

Very few of the items that I train my hawks to do involves a cue, but perhaps I should put it on cue anyway.

I have a project to train some delightfully irreverant behaviors to a goshawk sometime.

I would suspect that most of them have a cue, you just don't realize it. Animals are pretty good at finding one, it's humans that miss the details. Like the horse they all thought could do math, just looked at the person's eyes. The horse realized that it would stomp it's foot until the person changed their facial expression slightly and stopped, everyone cheered/reinforced the horse, horse loved it, and people thought the darn thing could actually solve equations.

[goshawkr]

I would suspect that most of them have a cue, you just don't realize it. Animals are pretty good at finding one, it's humans that miss the details. Like the horse they all thought could do math, just looked at the person's eyes. The horse realized that it would stomp it's foot until the person changed their facial expression slightly and stopped, everyone cheered/reinforced the horse, horse loved it, and people thought the darn thing could actually solve equations.

You know, Clever Haans could do math. It just used a human as a calculator. :D

[Mandragen]

back to the OP, I don't know if someone mentioned it but if they did I missed it. Ken Ramirez wrote a book Animal Training, it's more of a reference book than a read through, although I know many who have. He also has a website and does seminars, which I personally recommend. It's great information and he has an excellant way of relaying that so that it's easy to understand. His website has dates and locations.