BF Skinner, the man behind the best science-based animal training practiced today. He died in 1990, but, had he lived, today he would’ve turned 108. This is the first of several posts we’re putting on TheDogs today to honor him.
Nearly every aspect of good dog training boils down to figuring out how to control the consequences of a behavior in order to increase the chances that you’ll get the behaviors you want, and decrease the chance that you’ll get the behavior you don’t want. BF Skinner codified the laws of behavior and outlined exactly two kinds of punishment and two kinds of reinforcement: negative and positive.
Pretty much every action you make is either reinforced or punished; if the action is reinforced, it’s more likely to happen again; if it’s punished, it’s less likely to happen again. Sometimes, the reinforcement is hard to pinpoint, but it’s there. For instance, no one is handing me a bouquet of flowers every time I tie my sneakers, but I continue to tie a bow, just like I did yesterday and the day before, because the bow makes my shoe stay on my foot, which is rewarding. If, for some reason, the bow mechanism stopped working (laces break, gravity fails, whatever), then it’d decrease the chance that I’d continue wearing shoes that required bows. Or I’d just go barefoot. Which actually might be rewarding! If walking barefoot felt good, then that’d increase the chance that, the next time I were to face the choice between shoes and no shoes, I’d go shoeless. Unless I stepped on glass. That would be punishing, and would decrease the chance that I’d continue to walk around barefoot. I’d be sporting Velcro instead.
Skinner, genius though he was, was kind of bad with nomenclature and branding; his “negative” doesn’t mean bad–it just means taking something away. And “positive” doesn’t mean “good.” It just means adding something to a situation. But the pairing of the word “positive” with “punishment” is all kinds of confusing. It makes punishment sound good, when really it just means that something bad is being added to the equation. Like, say, a fist being added to your teeth.
In Burkhard Bilder’s excellent New Yorker piece on police dogs last month, he spelled out the quadrants of Skinnerian conditioning really nicely:
Skinner argued that it’s pointless to imagine what’s going on in an animal’s head. Better to treat its mind as a black box, closed and unknowable, with inputs that lead to predictable outputs. Skinner identified four ways to manipulate behavior: four buttons to push–positive reinforcement (“Good dog! Have a biscuit”), positive punishment (“Bad dog! Whack”), negative reinforcement (“Good dog! Now I’ll stop whacking you”), and negative punishment (“Bad dog! Give me back that biscuit”). Connect an action to an outcome and almost any behavior can be trained.
And that’s pretty much the foundation of any kind of dog training. Trainers who primarily use positive reinforcement call themselves clicker trainers. Or positive reinforcement dog trainers. Or +R trainers. See, we still haven’t fixed the wording issue. I’d say we should call ourselves Skinnerian trainers, except that I think that part of Skinner’s problem was his own last name. When I first started researching him, I found exactly one copy of one book about him in the Barnes and Nobles flagship store. There are definitely multiple reasons for his fall out of favor, but I can’t help but wonder if one of them is the fact that his name sounds like he’s about to rip someone’s flesh off. With the last name “Grossman,” I feel uniquely entitled to judge the ickiness of someone’s name. But I really think good things might have happened for Skinner if he’d just had just been called something else. Is there a reality show for rebranding dead scientists?
What if he called him Skinny. That’d make me a Skinny trainer. Yesssssss!
More on Skinner: