Comparing apples and oranges and coming up with magic beans

There's a quiet battle going on the travel world right now. It's a conflict between the people who believe in the so-called natural phenomenon called "gravity" such as high school physics teachers, and the people who believe in air travel, like Richard Branson, CEO of the Virgin companies.

Proponents of gravity believe that objects of mass attract one another and as a result, objects on or near the earth are irresistibly pulled toward it. Meanwhile, believers in air travel hold that one can enter a specially built craft and in defiance of gravity, fly in it to faraway destinations.

It appears that as more and more people use this "air travel" that gravity is failing our high school students.

No, you haven't accidentally found your way to the Onion. I saw this blog post over the weekend and the above sprang into my mind.

The author of that post compares learning theory to what he calls "alpha theory." He puts them on an equal footing in terms of scope and effectiveness, and then uses the "success" of alpha theory (which even he points out is based on a fallacy itself) as proof that learning theory is either wrong or incomplete.

This is sort of like pointing to aircraft as proof that gravitational theory is somehow flawed. (Except for the analogy to really be accurate, planes would either have to crash a lot more often or we wouldn't be sure how they worked.)

If you know anything about how planes fly, you know that as opposed to defying gravity, gravity actually plays a part in how they fly. There are other forces at play too, and to attempt to define everything a plane does in terms of gravity it to not fully grasp the situation.

In this way, operant conditioning (which is what the author of the blog post seems to be trying to question or refute) is one of the things that is at work when "alpha theory" is working (or not.)

Operant conditioning describes how the consequences of voluntary behaviors modify their frequency. Punishment makes a behavior occur less often. Reinforcement, more often.

Operant conditioning is observable, measurable, and reproducible. However, I guess if you use strict scientific terminology it is still a theory.

"Alpha" (or "pack") theory describes the relationship between two or more canids (it's also used with other species, but we're talking about wolves and dogs here.) As the author states, it's been heavily disputed (if not completely disproven,) in wolves.

Until we figure out some way to see or hear what dogs actually think, pack theory is not truly observable, measurable, or reproducible. I'm not sure if the idea even deserves the label "theory."

Comparing the two on an equal basis is apples and oranges. Maybe even apples and fructose. Even if pack theory still did have a place in science, it operates on a completely different level than operant conditioning - the interaction between two or more organisms versus how one organisms responds to stimuli. One actually fits quite neatly inside the other.

There is nothing in pack theory that negates or precludes operant conditioning. As a matter of fact, many of the training techniques show on the Dog Whisperer, who the author explicitly mentions and is also the world's most visible and successful proponent of pack theory, can be easily explained in terms of operant conditioning.

The problem with operant conditioning is one of scope. It's a one-dimensional view into a multi-dimensional system. Not everything can be defined in its terms, and I am not aware of any serious trainers, behavior consultants, or behaviorists that would try to do so. Beyond the apples and oranges comparison, this is what confuses me most about the blog entry; no knowledgeable trainer would try to solve a serious behavior problem with just operant conditioning. Anyone truly knowledgeable about learning theory and dog training/behavior consulting is aware of classical conditioning, as well as the specifics of canine behavior.

Another puzzling aspect of this blog post is the assertion that "dogs can also be trained without using any aspect of either theory." So puzzling that I'm not really sure how to respond. The fact that dogs can be trained without pack theory strikes me as entirely unsurprising since even the author himself disputes the theory.

The assertion that dogs can be trained without using operant conditioning is a bit odd considering the example linked in the "footnotes" of the post. The main gist is that by using praise the author believes he changed his dog's emotional state. That is classical conditioning, the very type of learning that is ignored in the referring post.

Unfortunately, this kind of either/or, right/wrong argument is all too common. When traveling through a forest it's admittedly easy to get wrapped up in seeds, bark, and leaves. Operant conditioning is a pretty "low-level" theory — the fact that it can be applied and observed in so many different species is proof of that —and making the sort error in scope that leads one to believe that you've discovered something new or different is very possible when you can't see the big picture.