Does anyone train positively without the treats?

If you're going to be using a food lure, it's important to fade it out quickly so that this doesn't become an issue.
I can often fade a food lure out after a few short sessions, switching to a non-food lure which then becomes my hand signal. My hand signal is paired with a verbal cue and is then faded out.

Below is a simplified example of what I do often do:

- Food lure, reward
- Lure, reward
- Hand signal, reward
- Verbal cue, hand signal, reward
- Verbal cue, reward
- Verbal cue, reward intermittently
- Verbal cue, variable rewards intermittently


I wouldn't think you would ALWAYS have to give a treat in order to get a response. The principles of learning theory are the same. My cats gather in the kitchen every time I open a can. More often than not, the can is NOT being opened for them so they are not always rewarded for gathering in the kitchen, yet they continue to gather in the kitchen on the chance that they might be rewarded. The sound of the can being opened is their cue to come. If I wanted to get a reliable recall using my voice I would simply say "Cats Come!" right before opening the can. I would do this each time I was to open a can. Soon the cats would begin to respond to my verbal cue because my verbal cue consistantly preceeds the sound of the can. I would no longer need the sound of the can to get them to come.

Having said that, 5 of my 7 cats have pretty speedy recalls. All I have to do is call their name and they come running almost every time and this is not even something I taught them intentionally. I used to joke that some of my cats had better recalls than my dogs.

Mocha couldn't care less about treats. It's the stuff animals that he wants. Missy will do anything in return for the good chase after the ball. Treats do not motivate her as much either.

So far, Muffin is the only one who love to eat, but she's only 9 weeks old I hope she doesn't grow to be as picky as her bro/sis when it comes to food.

I trained Mocha & Missy using mostly praises, and it seem to work for us. So I guess it is possilbe to train dogs without treats, but they do need to associate good behaviors with the good stuffs, whatever it is according to them.

If the praise becomes a predictor of non-punishment then it would actually be functioning as a conditioned negative reinforcer, right?

I had to think about that for a second ... yep.

I thought about it too....long and hard. LOL. NO fair throwing in trick questions. I think it's one of those grey areas if there is such a thing. I think at first it appears to be that....and maybe it is in essence. But here's what I decided. Maybe I'm wrong.


No. If a correction isn't happening at that moment, (no steady tight collar, no ear pinch) it is not being taken away. So, praise would not be a marker (or conditioned reinforcer) of having a correction (bad thing) taken away. There's too much of a time lag between the removal of something bad and the marker, (in your example, the praise) I think to be considered an actual marker signal.

As a predictor though.... as a historically or previously learned set of operants or behaviors and consequences.... if a dog is trained with a lot of stern punishment or collar yanks, praise tells the dog that he's not likely right now, to receive (have added) a collar correction, pain, discomfort etc. (positive punishment) Good things and bad things haven't ever happened simutaneously. So, praise is a pretty good thing. It has already been learned to indicate no punishment is going to happen. "Woopie!" But it is not indicating that a bad thing is going away. That would be negative reinforcement.

A conditioned negative reinforcer or bridging stimulus would be a word or some signal marker which communicated to the dog that the bad thing is about to stop.

Now, I'm not saying that praise is not a good reinforcer for all dogs or for all situations. I'm just saying (the above) and the fact that most people give attention, praise, love, cuddles, talking to a lot of the time anyhow. I do anyhow....lots of affection. I like it. So, when it comes to training sessions or what not, for the most part, I think a higher value reinforcer does a better job, especially with new things which need lots of reinforcement. It marks more precisely the behavior than praise which lasts longer so the timing is better.

I notice an improvement....definitely with my dogs, (after having trained using no treats or toys)although lots of times on our hikes I might forget to have treats on me and with behaviors or skills they know well and have been doing forever, I don't need them anymore, except, like I said, sometimes to keep them from regressing. If I want to teach something new, I definitely use treats and/or our favorite rope toy, depending on what it is we're working on. Perfecting the heel, I like tiny treats that I can pop in his mouth without breaking stride or the pace. For a nice, long down stay, a treat, a scratch behind the ears and a game of tug he loves.

Now you've got me thinking....
In doing a search to learn more about conditioned negative reinforcers I came across this article by Karen Pryor: Operant Conditioning and the Traditional Trainer

I'm slightly confused on this part:
Here is an example from the article:
I think this is sort of along the lines of what I was thinking....a signal to the dog that his behavior has just prevented an aversive. Then again, maybe this would only be true if the dog was so worried about being corrected that the word "good" gave him relief and caused him to repeat that behavior in the future.

Any thoughts on this article?

Is the behavior actually changing due to the negative reinforcer or due to the praise? I gotta look at this again. Behavior has to change to call it anything. What behavior are we looking at anyway? LOL. I need more coffee.

I think you're absolutely right. It seemed to me that it would have to be more directly connected, less of a time lag, but I guess the understanding that the dog is avoiding a punisher is what matters ultimately. Normally, we connect things immediately...as in the click and then treat. It would also have to depend on what the dog perceived when he heard the word, "good."

From Karen Pryor's book:

A negative reinforcer is any unpleasant event or stimulus, no matter how mild, that can be halted or avoided by changing one's behavior. A cow in a field with an electric fence touches her nose to the fence, feels the shock and steps back which stops the shock. She learns to avoid the shock by not touching the fence. While touching the fence has been punished, the behavior of avoiding the fence has been reinforced by a negative rather than positive reinforcer.

These things are actions on behavior. We have to think what behavior we're talking about. LOL.

So, he is moving into the correct position to avoid an unpleasant thing. That behavior (the moving) has changed due to a negative reinforcer, the all too typical and learned bad thing that happens when he lags.

If a dog routinely gets a lot of punishment in his training, this "good" can mean that no leash pop is going to occur. So.....even if it's not immediate, I see that it would be called a negative reinforcer.

If a dog gets rewarded consistantly when hearing the word, "good," then "good" becomes a conditioned positive reinforcer because a reward or reinforcer follows the word, "good."

So, it must depend on the dog's perception of the word, "good."

Here's another example in Karen Pryor's book: Some people find the smell of garlic appetizing and others find it offensive. The stimulus becomes a negative reinforcer only it if is perceived as unpleasant by the recipient and if the behavior is modified....shifing seats on the bus, away from a garlic-eater say...to remove the unpleasantness.

Here's something else:

Negative reinforcement can be used to shape behavior. As with positive reinforcment, the reinforcer must be contingent upon the behavior; one must cease "prodding" when the response is correct. Unfortunately, because the prodding, in whatever form results in a change in behavior, the behavior of the person doing the prodding may be positively reinforced, so that, as with punishming, the tendency to lay on with the aversives increases. Naggers, for example, may eventually get reults, and this is reinforcing to the nagger. So nagging escalates, sometimes so much that the nagger goes on nagging whetehr the desired response has occrred or not. Think of the mother in Portnoy's Complaint who complains, while her son is visiting, "We never see you!"

Very good article in that link. Thanks so much for sharing! I wish everyone would read it. So many good reasons not to use a lot of negative reinforcement. It creates avoidance and fear. Do you have her book, Don't Shoot the Dog? I read it and I see I need to read it again. Some of this gets muddled a little. I need to reinforce myself and read it again, along with Culture Clash. LOL.

I can see where I was a bit off on the praise thing. It makes more sense if the CNR is given during the behavior. Great post Doberluv!

Yes, I have her book too. I was just looking through it last night...its been a few years since I read it.

Well, I'm not so sure you were off. I think I was off. LOL. Is that called "balance or what?"

I would like to learn to train without having to always have a treat... my dog is very food oriented, but he also responds to attention, toys, kisses, belly rubs, even an enthusiastic "Good boy!"

In fact, he does pretty well on loose-leash walking and I rarely have treats while walking, he's learned it all with "Good boy!" whenever he was next to me.

The reason I'd like to learn how to use less treats when training is because I have a limited amount of repetitions I can work: He is on a diet of 1/2 cup kibble per day. That gives me 50 pieces of kibble or so to use throughout the day -- I can easily go through that much on one walk (we're working on being polite when other people are around, not barking or growling. He's doing really well!)

I also add a few treats throughout the day as well, I do a couple recall exercises and reward him for being quiet at work when I leave my desk for a few minutes.

But the vet says he's overweight so I have to be careful. It's really frustrating when he's already eaten all his food for the day and I want to train something. I can do two or three reps with a bit of treat (really small pieces) but I wish I had more to work with.

That's the hard thing about having a little dog!
I can just imagine how much I could teach him with 4 cups of kibble to work with!!