raining a hawk for falconry is an incredible process and awe inspiring to even the experienced. Apprentice birds are typically trapped from the wild after they have started supporting themselves but before their first moult. These birds are therefore already hunting and the falconer is not there, at first, to train the bird to hunt. She has already been doing that, although with varying degrees of success. The training initially is to acclimate the bird to accept the falconer's presence, both as her keeper and as her companion in the field. Because most of these birds are not social, cooperative hunting is a tentative relationship and a delicate one.
Traditionally the process of acclimating the bird is called manning. It consists of first getting a bird over her fears and then working to get her to allow you to do more with her.
One way to man a bird is to flood her senses with her new environment. Traditional methods such as waking take advantage of a bird's state of shock after being trapped. The bird's mind is in shock as she is prepared to be mauled or eaten. She literally freezes and the falconer can take advantage of this. The combination of the falconer holding the bird and letting all of the normal experiences of what will be her new life wash over her acclimates a bird very fast to the household, TV, dogs, children, other people, and more. Many falconers practically hold a party to ensure that this period of time is fully taken advantage of. The bird is not allowed to sleep for an extended period of time, usually more than 24 hours as the sleep deprivation helps the manning process. The bird's system is also left empty for a day or so allowing her to come down to her flying weight - that weight at which she is most motivated to be hunting. Many birds that have gone through this process are rock steady around any number of unusual objects or experiences. However, there is also a very common occurrence where the bird will develop a deep seated fear of one or several things from this. This could be a fear of hands or hats or red dogs. It's very difficult to predict how some of the fear will manifest itself and most falconers will just manage the fear as addressing it can be very complex.
The opposite approach is to deprive her senses of her new environment and introduce pieces slowly. The Japanese and some European falconers have traditionally used another method whereby the bird is kept without the ability to see. It may be hooded or in a dark room, but whatever the method, the handler controls the speed at which her new environment is introduced. The birds are so visually focused that if they do not see anything scary, then there is nothing scary to be afraid of. Gradually the light in the room is raised allowing her to take in the objects and the people slowly. The advantage of this system is that the bird has very little stress either on her system or from bating. There is almost no fear as nothing scary has ever happened to her. Many birds who were trained by this method, or variants of it, can be flown at much higher weights and with less fear of being lost. The disadvantage of this method is that it can take much longer and can take more patience on the part of a falconer. This method can take a month or so to be out in the field and hunting with dependability.
Both ways work, as do many others. Each has a drawback in some people's eyes.
Beyond manning is the process of training a bird. A well manned bird has no fear, but that does not mean he is well trained. This encompasses everything from basic husbandry (to have the bird allow you to clean her talons, for example), to conditioning the bird, to getting the bird turned on to new quarry.
The basic philosophy that I ascribe to is operant conditioning. This was introduced to falconry by Steve Layman who has been using it and experimenting with the techniques on various raptors for decades. Operant conditioning (OPC) is commonly known as clicker training in the equestrian and dog training circles. It is a training method using positive reinforcement to achieve results.
Classical conditioning is often described as stimulus - response. Operant conditioning is response - stimulus, or stimulus - response - stimulus. An example of the response - stimulus chain might be a trainer watching a dog in a room. As soon as the dog performs an action that the trainer wants repeated, a treat is produced. If the dog is wandering around in the room, sits down and a steak appears, he will try sitting down again to see if another steak appears. An example of the three-step chain might be asking a dog to sit (stimulus is the word, "Sit!"), the dog sitting (response is the dog's action of sitting down), followed by the trainer handing the dog a treat (stimulus of the dog getting what it wanted). The second stimulus is a reinforcement.
Reinforcement makes a behavior more likely - the behavior is reinforced. In the above case, the reinforcement was positive because the dog was trying to get that same stimulus to occur again. The reinforcement was a positive to the dog.
A negative reinforcement is when a negative is removed to produce a behavior - the dog is trying to avoid a stimulus. This can be seen with dogs. If you push down on a dog's hind quarters, most dogs will resist slightly. The instant that the dog stops resisting and begins to sit down, the pressure is taken off the hind quarters. In that case the removal of the pressure is a reward to the dog for sitting down.
All the reinforcers described above are primary reinforcers. Primary reinforcers are those that biologically are desired by the animal. Food is an excellent example of this and a frequent primary reinforcer used by trainers. However food is not always convenient or desirable to use. It is not very precise and can become too dissociated from the exact action that is to be rewarded. If a dressage rider wanted to reward her horse for a particular movement, it could take a moment or two to stop a horse, reach down, grab a sugar cube and get it to the horse's mouth. For several reasons, a secondary reinforcer is used.
A secondary reinforcer, or conditioned reinforcer (CR), is one which must be learned, the animal is conditioned to it. If your three year old son did something good, would he prefer a lollipop or a small green piece of paper? And yet every day many of us go to work for small green pieces of paper as we know they are good - we have learned about them and are conditioned to work for them. Similarly if we want to use a secondary reinforcer we must teach the animal that it is good and to work for that as much as a primary reinforcer. The reinforcer may be anything. The popularity of clicker training has made the small "clickers" highly popular. However some animals are afraid of the sharp sound and, more importantly to me, I cannot manage all the individual pieces of equipment along with a clicker in my hand. The positive reinforcement could be a flash of light, a hand movement, a word spoken by the trainer, or any other event. Sounds are very easy to produce and the animal doesn't have to be looking in any particular direction in order to be rewarded. However, the sound must be consistent each time so that it is recognizable. An effective CR is one which is specific, precise, and recognized. I give a short whistle as my CR.
The process of creating a conditioned reinforcer is called conditioning. Some texts refer to this as "charging up" the secondary reinforcer. In dog training this may be done by clicking a clicker as the dog eats or clicking while treating the dog. The dog learns to associate the clicks with the very positive event of eating. Initially when the dog is rewarded he will receive a secondary reinforcer and a primary reinforcer at the same time - a click and a treat. As the dog becomes conditioned he will only receive the click with treats at various intervals or schedules.
It would be highly effective for your boss to sit next to you all day and hand you a $10 bill for every correct action you did. However, he only has so many $10 bills to hand out and he could run out at noon if you got very good at what you were doing. So he would need a way to tell you which actions were correct and keep you motivated without it costing him a lot. This is where scheduling comes in. A primary or secondary reinforcer may be given after every correct action. This is called continuous reinforcement. However, there may only be a finite amount of reward (treats, money, etc), or the trainer may want to start shaping the behavior, or the animal may start trying to do less work. To get more correct behaviors from an animal without using as many reinforcers, you could reward every 3rd behavior, or every 4th behavior. This is a fixed ratio schedule - the interval between reinforcements is fixed. This will stretch out the training, but then the animal learns that immediately following a reward, there will be a stretch without a reward and he may perform accordingly. In order to produce the most dependable behaviors, a variable reward schedule is best. Layman describes this as the slot machine - it can keep a person fixed in place for hours while they wait for the next pay out. Likewise, it will maintain the same response for a very long time.
Punishment makes a behavior less likely. The word "punishment" has so much baggage and conjures up so many thoughts that we must be precise about what a punishment is. For an animal like a dog, getting petted for sitting down is a positive reinforcement - he wants to be petted and will sit down for that. If that same dog has started barking and his owner leaves the room, he has just been punished. The owner's presence was a positive for him and a desired thing. Removing that was a punishment to the dog. The dog may take a while to connect his barking with his owner leaving him, but the event was a punishment. Many people, and even trainers, confuse the concepts of negative reinforcement and punishment. The most simple way of thinking of this is that a negative reinforcement is actually desired by the animal while a punishment is not. If a child misbehaves he may go to bed without dessert - he has been punished for the behavior.
Shaping behavior is what trainers do to get a desired result out of an animal. However, if I want to shape a behavior, I need to select for a certain part of a current behavior to modify it into the behavior I want. The way I describe this is through a scenario. If I am training my bird to do fast jump-ups, then there are several components - she can jump up to the fist quickly, she can jump high or powerfully, or she may throw out a behavior that I want to select for, like jumping and turning at the same time. If I start to reward only those times that she leaps quickly after she has landed on the perch, then she will start to jump faster spending less time down on the perch. I may also select for the turn-arounds so that she is doing extra work to jump fast and turn around in mid air.
In order to be effective and start using operant conditioning, it doesn't necessitate using fancy language or even understanding the concepts fully. You're not going to break your bird (or any animal) by starting to use it. The terminology is there for higher level discussions so that people can be precise in their communication about what they are doing and why they are doing it. Don't be put off by the language and science. It is a very powerful methodology and yet can be applied at some level by anyone. Applied properly this methodology will always work.
For more information on Operant Conditioning and falconry, Karen Pryor's Hawk Chalk article is available here: http://www.clickertraining.com/node/282
Training your first longwing: http://www.americanfalconry.com/FirstLongwing.html
Training a hawk with OPC: http://www.youtube.com/watch?v=8k7B-JGowL4