Thursday, November 09, 2017

The Truth About Consequences

Operant conditioning was featured in the television sit-com The Big Bang Theory, above, but of course it finds its way into our news cycle every day (see today's news headlines appended at bottom of this post).  

Notice what Sheldon says about the relative speed of using pleasant rewards alone, as compared to a more balanced approach that pairs rewards with less pleasant consequences. 

Of course in the real world things are a little less simple than the scenario presented here, and many issues have to be factored in, such as what behavior is being shaped, how long it has been going on, whether it is neutral, or whether it is strongly self-rewarding.

For example, would chocolate alone have been enough to prevent Leonard and Penny from having a sexual liaison?  Don't count on it!  

Would it be possible to "train a different behavior" for Leonard and Penny and have it reliably stop all geek love?  Don't count on it!

What if the behavior that Penny was presenting was not just annoying -- it was extremely dangerous to her, and a single wayward incident could kill her?   Would it be OK to engage in a little aversion then?

This last question is not a contrived situation, but one which Karen Pryor glosses over in her book Lads Before the Wind: Diary of a Dolphin Trainer.

The porpoises and whale themselves, in their quests for entertainment, often created problems. One summer a fashion developed in the training tanks (I think Keiki started it) for leaning out over the tank wall and seeing how far you could balance without falling out. Several animals might be teetering on the tank edge at one time, and sometimes one or another did fall out. Nothing much happened to them, except maybe a cut or a scrape from the gravel around the tanks; but of course we had to run and pick them up and put them back in. Not a serious problem, if the animal that fell out was small, but if it was a 400-pound adult bottlenose, you had to find four strong people to get him back, and when it happened over and over again, the people got cross. We feared too, that some animal would fall out at night or when no one was around and dry out, overheat, and die. We yelled at the porpoises, and rushed over and pushed them back in when we saw them teetering, but that just seemed to add to the enjoyment of what I'm sure the porpoises thought of as a hilariously funny game. Fortunately they eventually tired of it by themselves.

Yes,  fortunately, they eventually tired of it by themselves.   


One has to wonder, however, if perhaps a little aversive natural consequences had something to do with it. 

For instance, what about those scrapes and cuts?  If a trainer did those that would be horrible, but if the animal did it to itself and changed its behavior by itself, can we then say "fortunately"?  

And what if the trainers were a bit late or a bit slow to get a porpoise back into the tank?  Did the animals get both hot and uncomfortable?  If a trainer did that on purpose, of course, that would be horrible and cruel, but if it was simply "one of those things" and the animal learned and changed its behavior, then can we use the word "fortunately" again?

And what if we have the exact same situation, but instead of a porpoise, it's a dog climbing out of its kennel?   Freedom for a dog (especially an intact male that can smell a female in heat) is a very self-rewarding behavior, and most dog owners have had a dog climb out of a kennel as a result.  The age old solution is a fabric kennel cover and a hot wire.  Is it OK for an owner to put in a hot wire to create an "unnatural" consequence that, to the dog, will seem as natural as any other?  If not, why not?  If so, why so?   Does it matter that the dog may die if it gets out even once?




Sue said...

Yep, training animals is never as simple as any approach seems at first glance. I remember the lesson from Bob Bailey in Chicken Camp: Classical conditioning always trumps Operant conditioning. Or, "Pavlov is always on your shoulder." I am a dog (and chicken) trainer who uses clicker training a lot to shape behaviors--and that's an axiom that is constantly making itself obvious. :) The other important factor is outlined in the Breland's article, "The Misbehavior of Organisms," on how instinctive behaviors can affect operant conditioning. Withhold the reinforcer too long and you get a chicken who scratches (dances!) and can't attend to the cue.

Tese are hard concepts to get across to students who conclude that operant conditioning "doesn't work" when in fact they simply failed to manage the environments or understand the relative value of their reinforcer vs. reinforcments and aversives in the immediate situation.

As for preventing dangerous escape behaviors--I do advocate the judicious use of a hot wire. It is a cheap and effective solution to a dangerous (and often instant) habit, and it turned my escape-artist Kangal Dog into a fence-respecting lady. After she stopped looking at and calculating that wire atop the fence (about 6 months) I turned it off and never needed it again. She forgot the thrill of escape, it seems. Mind you: she was a tough dog, so a couple of shocks did not turn her into an emotional muddle or make her run howling at the sight of a wire; aversives have different effects on different animals and can have unintended consequences. Not something to experiment with lightly, IMO. Which is why I have a problem with some of Millan's methods--the disclaimer at the end of the show is not enough to deter people from thinking that kicks, jerks, and flooding will mamgically work for their dog like it did on an edited TV show. I see the results in my classes, with the angry frustrated owners who resort to punishment as a first option, and can't get why their frightened, stressed-out dogs avoid them while performing like happy learning machines for me. Often, they are so vested in their beliefs about "dominance" and "being the pack leader" that they drop out before they ever learn another way.

PBurns said...

For more on the Breland's and the "Misbehavior of Organisms" see >>

Bob Bailey and others are actually reinforcing the natural behavior of chickens and pigeons. If you try to get a chicken to NOT scratch and to NOT peck -- i.e. you just want it to stand there -- you will find it cannot do it for very long. The code is the code, and it's best to work with the code rather than against it.

Every dog is different, but most dogs are not fragile eggs which is why they have done so well with humans for so long.

Sue said...

True that a chicken will revert to "fixed action patterns" (I think that's the term) under stress (like when a learned behavior fails to "work" repeatedly )--and the Brelands found that very easy to reinforce and put on cue for all those state fair chickens-in-boxes. WIth cats, it's the circling and rubbing against your legs. But chickens and cats can learn to stand still or carry out a chain of behaviors for a rather long time. It's just that you have to shape it carefully and not stress the animal so that it reverts to those FAPs--which can really mess up your training goals. That's one of the hard lessons that we learned in chicken camp. Timing, criteria, and rate of reinforcement have to be carefully planned and applied to get animals to do complex behaviors that don't come naturally.

Gina said...

Wouldn't it be swell if people actually believed their own lying eyes? Used common sense instead of making training into a religious cult?

GreenGrrl said...

Yep. Some dogs are soft and some are hard. Different methods for each dog. My little poodle/terrier rescue mutt is a pretty hard-headed little bitch. She's pretty good with verbal and hand commands now, but sometimes I have to give her a little nudge to remind her on how to act civilized in public. She basically looks like a muppet, but this little rescue muppet likes to follow her nose and dig stuff out (also proudly presents me with the mice/lizards/etc). She's also extremely protective of me. She'd be a good reminder that just because you adopt a little dog with fluffy, curly hair, doesn't mean you might not get a lot of dog with that. She's my first small dog and I have to say, she's more challenging than the GSDs and Pit Bulls, I grew up with. Albeit, less challenging than the Border Collie I had. Loved that dog, but to put it this way, when I got him, his trainer looked at his papers and said, "you've got yourself a fine pup, but he comes from a line of hard dogs."

Unknown said...

This is why children of veterinary staff are very well behaved.

LRM said...

Doubt the Segel weight loss effort was especially effective. Doesn't matter in the slightest...he's still wonderfully original (recently saw 'Forgetting Sarah Marshall' again + Russell Brand... hilarious )