Why Corporate Training May Be Less Effective Than Bird Training

Flying from Newark to Hong Kong, I indulged myself in catching up on the New Yorker.  This issue (May 12, 2008) was a double-delight, and offered inadvertent insights on corporate training.

I always welcome a new Malcolm Gladwell article—"In the Air: Who Says Big Ideas are Rare?" is on the nature of creativity,  and the frequency with which major scientific inventions are discovered simultaneously by many. (Not just Leibniz and Spinoza with calculus, five people came up with the steamboat, Gladwell says, and 9 invented the telescope at the same time).

Gladwell draws the implication that “genius” can be socially engineered by linking up smart people in a focused manner. Living proof is former Microsoft Chief Technology Officer Nathan Myhrvold’s company Intellectual Ventures, which has generated thousands of patentable ideas.

But you can’t over-engineer creativity. Randomness must be given its due. A starting problem and some pre-work seems to be important. Beyond that, the problems that end up being solved may vary radically from the original agenda. People must do their thing for group genius to take hold. Beyond agenda, it’s the confluence of smart people, not the proscribed process, that gets results.

But Gladwell’s piece is not all. The same issue has the amazing story of Alex the parrot, trained for 30 years by the equally impressive scientist, Irene Pepperberg.  Margaret Talbot’s article, "Birdbrain: The Woman Behind the World’s Chattiest Parrots," begins by reprising our historical view of animal intelligence.

Descartes, Talbot writes, posited a hard line between human intelligence and the “sub-human” view of animals. Darwin threw us back toward anthropomorphism. With B. F. Skinner, we returned to the Cartesian view of animals—but included people as essentially animalistic.

When Alex the parrot came on the scene, he confounded the animal behaviorists. He would initiate conversations, make appropriate comments, ask for things. “Sit on your knee?” You let him. “Scratch my head.” You do. Alex would ask, “You want food?” No. “You want drink?” No. “Then what do you want?”

Call that what you want, Alex was no birdbrain. Alex took conversations past where we’ve gone with chimps and dolphins. But here’s what drove the Skinnerians batty. Alex was not trained through stimulus and reward; he was trained through social interaction. Given classic “push the button get the pellet” training, he learned nothing. But placed in a social setting, competing and interacting with another (bird or human) for a trainer’s social attention, Alex learned rapidly. The author says:

As it happens, the model/rival method may have some utility for another species—humans. Diane Sherman, who works with autistic children in Monterey, California, has had some preliminary success in encourage speech in her clients using Pepperberg’s protocol. In an article published in The International Journal of Comparative Psychology, Sherman and Pepperberg say that, in two studies of children in Shermans private practice, [Pepperberg’s] model/rival method led to “significant gains” in the children’s “communication and social interaction with peers and adults.”

What does this have to do with corporate training, you ask?  At the risk of offending a few valued readers, let me make an overly broad generalization.  Corporate training is one example of a corporate trend toward breaking things down into parts and parceling them out.  That’s the pattern of outsourcing, of decentralization, of measuring outcomes, of securitization and disintermediation in the mortgage industry

These corporate trends are not all bad.  "Breaking it down and parceling it out" sounds like time-honored specialization and division of labor.  Division of labor is a recipe for efficiency. But we often lose sight of a holistic perspective and objectives. In training, this often means choosing training over education, sacrificing learning in the pursuit of behavior. Judging by outcomes, curriculum designers long ago ditched John Dewey in favor of B.F. Skinner.

In the corporate training world, every program has cascading objectives—overall, daily, and module-specific. Each “objective” is typically phrased in terms of behaviors: “participants will learn the skills and behaviors associated with….”

Can you imagine going to church or synagogue and being given a "spiritual inventory" from which to develop a "gap assessment" and a "needs analysis" followed by a customized program?

"In this class, sinners will learn the behaviors associated with the sixth through tenth commandments; participants will learn not to kill, and will no longer covet… Pre-requisite course: commandments 1 through 5."


Myhrvold’s lesson—underscored by Gladwell—is that objectives work at the level of agenda, participants, and pre-reading. Programming creativity below that level of specificity just kills it. Yet a modern trainer would be shot for proposing a program as open-ended as Myhrvold’s. In module design, an exercise yielding such varied outcomes would be tightened, trimmed, focused so that it produces replicable, predictable results—eerily called “learnings.”

The story of Alex the parrot tells us the same thing. Put me in a cage, appeal to my lowest level—peanuts and M&Ms—but you’ll get nothing from me unless you keep me starving. Give me the richness of social stimulation, give me some control over what I learn and how, and I will astound you with what “even” a bird can learn.

Talbot’s article has this poignant section about a famous research chimpanzee named Nim Chimpsky who had learned sign language:

After Nim had been retired to a Texas ranch where most of the employees didn’t know sign language, he continued to sign. When Bob Ingersoll, an old teacher of Nim’s, came to visit him the chimp, who was in a cage at the time, eagerly signed “Bob,” “out,” and “key.”

Even a chimp doesn’t like being processed.

3 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *