Friday, November 4, 2011

Ken Ramirez: The Path to Enlightened Training

It's a really awesome time to be a dog trainer. There's so much new information coming out to improve not only behavior and training but also our relationship with dogs. Ken Ramirez isn't just a dog trainer. He's the executive vice-president of animal collections and animal training at the Shedd Aquarium in Chicago. He's in charge of about 3,200 animals (and people think I have a lot of critters), and he's trained not only marine mammals, but also snapping turtles, sharks, and komodo dragons. Ken says learning principles are the same across all species. "You can train a Harvard graduate the same way you teach an earthworm," Ken tells us.

Well, he should know.



On Sunday, Ken started by talking to us about two types of reinforces: primary and secondary. A primary reinforcer is "inherently reinforcing" like food, water, or sex. Ken also puts the "drives" into this category, so play, prey, social interaction, and so on. However, he adds that while the drives themselves are primary reinforcers, but the object of the drive must be learned.   

Huh, what?

Okay, take play drive: your dog has a natural desire to play. But he or she has to learn to play with a toy - the object of the drive. I personally think that play drive is fascinating all in itself. Play in predators is usually practice for hunting. So the same behaviors you see in prey drive tend to come out in play. Hunting in wolves is broken down into six parts: orient -> eye -> stalk -> chase -> grab-bite -> kill-bite (Coppinger, Dogs). Selective breeding of dogs has created exaggerations in the hunting pattern. Border collies, for example, have been bred with an exaggerated eye and stalk but little to no kill-bite. By knowing which of your dog's hunting behaviors have been exaggerated (note I said "your dog" not "your dog's breed"), you can better pick out an effective toy for training and engaging play drive.

So, let's look at Rubi. She a terrier, so you'd think grab-bite would be a big one for her. She does love to play tug, but more than that, she loves to chase. Squirrels, anyone? She'll drop any tug for the chance to chase her big red ball. Unfortunately, throwing a ball is not the greatest real life reward for training. It works, but once the ball is out of my hands, I have no control of it. Plus we don't always have the space for me to throw a ball.



Here's where Ken's method for teaching secondary reinforcers may come in handy. A secondary reinforcer "acquires its reinforcing value through association with primary reinforcers." So you can teach a dog to enjoy a secondary reinforcer of your choosing by associating it with a primary reinforcer. Ken uses the example of pairing food (primary) with clapping your hands (secondary). But I'd like to try increasing Rubi's desire for a tug toy. This has the benefit of engaging two primary reinforcers: play and food. I feel like this should increase the strength of the secondary reinforcer (aka, the tug toy), but maybe I'm just American and think that more is better.

Anyway, according to Ken's method, we first train the tug by pairing it with food. So, present tug, tug tug, food. This has the advantage of teaching the dog to release the tug  - always tricky - by having them drop it to get food. Lather, rinse, repeat through several sessions until the dog lights up when they see the tug. Then ask for a simple, well established behavior, mark the behavior, and tug. Do this a maximum of three times in a training session. Once the dog is working well for for the tug, ask for a harder, well established behavior. Work your way up until you're asking for harder, newer behaviors. Once you've generalized your new tug to all behaviors, then you can start gradually increasing the frequency of use.



The hard part about secondary reinforcers is that, well, they're less reinforcing than the primary reinforcer. This means that you should never use more secondary reinforcers in a training session that primary reinforcers. If you do, your dog will slowly become less motivated to work with you because the reinforcement just isn't there. Ken states a 20/80 ratio of secondary to primary reinforcers is about right. You will also need to regularly pair your secondary reinforcer with a primary reinforcer in order to maintain its strength. Ken also says to avoid using the same secondary reinforcer twice in a row.

So how do you decide what reinforcer is best for you and your dog? Here's the big R again: relationship. Certain rewards are going to more effective in different situations. Knowing and understanding your dog is the best way to figure out what is going to be most motivating for him or her in any environment. Not to mention that secondary reinforcers are useless without a personal history.

For example, "good" is a verbal secondary reinforcer in my house. If Rubi and I passed a dog on a walk, you can bet I wouldn't be telling her "good" - I'd be shoving food in her mouth as fast as she'd take it. But on the same walk, passing the same dog, with Piper Ann, a "good" would more than suffice. But if I were walking someone else's dog, I wouldn't even think of using the word "good"; to that dog, it's just another meaningless word. There's no history there. No real relationship. And relationship as we hear time and time again, is one of the most important parts of having a dog.

9 comments:

  1. I'm confused. I see how pairing the tug with food will increase the value of the tug itself (the object), but don't you want to increase the value of tuggING (the action)? If the tug is just a means to an end the true value is in the food, right? Why not just do drive building? If she tugs at all, you should be able to increase that drive with tug alone. Right? Am I missing something?

    ReplyDelete
  2. According to Ken, pairing the behavior with food should increase the value of tugging. I'm not presenting her with the object and then rewarding her (although I could and it would still be a secondary reinforcer). I'm asking her to tug, then pairing it with food. It's not the object, it's the behavior. Thus far, driving building with the tug has met with mediocre results. At worse, she gets over stimulated and redirects. I'm hoping that a different tactic will yield slightly different results.

    ReplyDelete
  3. I get how you can reward the behavior of tugging with food, but I still don't think that's the ideal way to do it.

    In agility I've seen a lot of dogs who do the whole tug to get food thing and they never have that same intensity as those who tug just to tug, even with dogs who had to be "trained" to love tugging. With dogs who don't naturally love to tug you can certainly get that tugging behavior FASTER by pairing it with food, but it really becomes just another behavior the dog does rather than trying to make the tugging itself valuable. Of course in order to tug for tugging's sake you really have to work hard at it, knowing how bring out the dog's drive by "properly" playing with the toy, knowing how long to do it, etc. And for some people and some dogs the tug for food is certainly an easier and more practical option. I just have never seen it end up with that same intensity in their eyes even if their actions are the same.

    ReplyDelete
  4. Anyways, with Lance I don't do a lot of tugging as a reward. I've slowly built it up at home and he seems to really get into it at times. But in most other places food is the easiest and most rewarding for him. But I have managed to turn the db, articles, and glove into a little game by just having him chase it without actually letting him bite. He LOVES it!!

    ReplyDelete
  5. Hmm, maybe I should rewrite this for clarity's sake. The idea was secondary reinforcers, not tugging per se.

    Since she's getting over stimulated with tug, and introducing food into the game makes it less intense, isn't that a good thing? Thoughts?

    ReplyDelete
  6. I get what you're saying, and I guess it is the action--but only in the way that the act of sitting on cue is a secondary reinforcer.

    If the idea is to have a secondary reinforcer, why not just use a cue or something like that? Not that you shouldn't use tug the way you are. It just seems like a lot of work for just a secondary reinforcer. I guess the difference is that drive building makes it a primary reinforcer. But drive building can be tough. I'd work through the overstimulation by doing off switch work and build drive with short sessions and dismissal.

    ReplyDelete
  7. I plan to work in some other secondary reinforcers as well, tug was just the example I used for the blog. Whoops, bad example. Ken used clapping as an example for secondary reinforcers, and I probably would have been clearer if I'd stuck to that.

    I'm taking Ron's Bitework for Behavior class with B starting next week, and I'm hoping the more individualized training will help. I know I'm not using B's drive as effectively as I could, but I'm not sure where to streamline things. I didn't have this trouble with drive building for Piper or Allister (or Maus, for that matter, although he's the least drivey dog I've ever owned), so there must be something not quote right with how we're operating as a team.

    The rile/recovery and dismissal have been great for her attention, but she still gets over stimulated. Hmmmm . . .

    ReplyDelete
  8. I'll be interested to hear how the bitework class goes! Make sure to post about it!

    ReplyDelete
  9. No I totally get the gist of your post! I just don't think there's a reason you have to settle for tug being a secondary reinforcer when it's so linked to a direct drive.

    That being said it doesn't sound like you're using food in order to pair tugging with a primary. It sounds as though tugging is already a primary reinforcer and you're just trying to interrupt the tugging with something she can actually respond to when in that state, eat food.

    ReplyDelete