Very good thread!
I have been training dogs for over 15 years. I started Kohler method. It worked. Better with some dogs than others. It didn't teach me as much about the dogs, their learning process, and wasn't rewarding or fun for me or the dogs.
I searched for other methods to incorporate to make training more fun and interesting. I learned about "drive" training. It was more fun, it worked. Better with some dogs than others. Produced great results for some things, difficult to control in others. The 2 methods enhanced each other as my skills improved and I understood both more clearly.
A little over 10 years ago I discovered "marker" training (also referred to as clicker training). I find this to be a fun, rewarding training method which works with any dog, in most disciplines, and has taught me more about behavior and the learning process than almost anything.
When you learn how to use several methods well, and fairly, they can enhance each other. I can work in "drive", "mark" correct behavior, as well as "correct" undesirable behavior.
It is a constant learning process and no two dogs I have trained have ever needed the exact same process. They are all individuals, learn differently and need different things. Its important to learn and understand learning theory and understand and have the ability to apply many techniques and know what will work best with a particular dog in a given situation.
But I DO LOVE "marker" training!
Although I understand the premack principal I find it difficult to compare these types of laboratory studies to real life application.
When I started "clicker" training I really thought that the dolphin trainers were really something to aspire to. But it was pointed out to me that these animals were basically confined to a concrete "skinner box".
Hardly comparable to the same type of training in a normal, distractible environment. Its easy to have stimulus control when you have complete control over the entire environment... no new sights sounds or smells. It is a challenge to still have the same control of our dogs with all of the constant changes in their environment... wet grass, chickens, and all... and still have them free to investigate these new things. They can have so much more environmental experiences if they are reliable.
Similar to the limited knowledge of canine behavior obtained by studying dogs or captive wolves, was a kind of a handicapped version of what was learned when studying wolves in their natural environment became more plausible...
...drinking motivated by running, and running motivated by drinking, are kind of stupid learning observations when they are the only options for thirst and exercise in their environment. Who knows what kind of USEFULL learning theory could be studied if that experiment hadn't been so utterly simplistic. I don't know what real knowledge was gained from that, than from Skinner boxes.
None of my dogs are perfect, and each of their training is a constant learning process (more for me than them) and always a work in progress.