Someone asked me a fabulous question the other day. “Define change. Why is it so hard to change?” I responded, tongue-in-cheek, that “change is the only constant.” Upon reflection I think there needs to be some expansion of the question. As always, I am drawn back to Bateson’s learning model, which offers some clues.

In Bateson’s model, 0^th order learning indicates a simple, instinctual stimulus-response package. There is no change. The stimulus always produces the same response.

1^st order learning means that the stimulus is replaced to produce the same response for a new stimulus, or a new response is learned to replace the response to the same stimulus. Pavlov’s experiments are the best-known examples of this. Rote learning is one way to accomplish this. This doesn’t mean the change is painless, but it can be accomplished with brute force repetition and not much thought, reflection or systemic thinking.

2^nd order, or deutero-learning involves “learning to learn,” a kind of meta-learning. Here the change is much more drastic and in the context of training, education and learning in the organization, this is mainly what we are talking about when we talk about change. The issue is that learned responses tend to be grouped into patterns. This is good, as it allows us to reuse 1^st order learnings. However, we sometimes attempt to tackle unfamiliar situations by applying one or more learned patterns that may not be appropriate. On TOD this morning, there’s a comment thread on a discussion about Cyclone Gonu entering the Gulf of Oman. A cyclone has never entered the Gulf of Oman since records have been kept, so the predictions some are making are based on trying to map some other experiences to a novel situation.

The pattern formation process is really at the root of the difficulty. The ability to group similars into patterns is extremely useful, advantageous and efficient in terms of managing 1^st order learning. The ability to change at this level then, is predicated on the ability to suspend pattern formation in order to see difference. As Tufte says, information is “a difference that makes a difference.”

3^rd order learning involves meta-patterns and paradigm shifts and Bateson rightly points out why this is so rare. Suspending an individual pattern or grouping is difficult; suspending a pattern of patterns nearly impossible. The whole pattern-formation process must be reconsidered and reformulated along other, previously subordinate, lines.