In a word, learning is decontextualized. We break ideas down into tiny pieces that bear no relation to the whole. We give students a brick of information, followed by another brick, followed by another brick, until they are graduated, at which point we assume they have a house. What they have is a pile of bricks, and they don't have it for long.

Some who support [more] coercive strategies assume that children will run wild if they are not controlled. However, the children for whom this is true typically turn out to be those accustomed to being controlled— those who are not trusted, given explanations, encouraged to think for themselves, helped to develop and internalize good values, and so on. Control breeds the need for more control, which is used to justify the use of control.

Non-cooperative approaches, by contrast, almost always involve duplication of effort, since someone working independently must spend time and skills on problems that already have been encountered and overcome by someone else. A technical hitch, for example, is more likely to be solved quickly and imaginatively if scientists (including scientists from different countries) pool their talents rather than compete against one another.

If a child is off-task...mayb e the problem is not the child...maybe it's the task.

We can't value only what is easy to measure; measurable outcomes may be the least important results of learning.

Very few things are as dangerous as a bunch of incentive-driven individuals trying to play it safe.

The race to win turns us all into losers.