Bickhard, M. H., & Campbell, R. L. (1996). Developmental aspects of expertise: Rationality and generalization. Journal of Experimental & Theoretical Artificial Intelligence, 8(3-4), 399–417. doi:10.1080/095281396147393
Cited by Carl and Marlene when talking about promisingness in QWERTY paper.
An expert is someone who knows some of the worst mistakes that can be made in his subject, and how to avoid them. Werner Heisenberg
We will use the term critical principles to cover any sort of knowledge of types of errors (Bickhard 1991a, forthcoming). Critical principles are not positive knowledge -knowledge of what works or what is true. They are negative knowledge -knowledge of what to avoid.
Critical knowledge must be constructed, just like any other knowledge. And critical knowledge is subject to reflection, just like any other knowledge.
Still, the construction of knowledge is progressive–in a negative sense. Culturally or individually, progress means coming to avoid more kinds of errors.
Positive knowledge is often overthrown as new sorts of error are discovered ; negative knowledge is more likely to survive (though its survival is by no means certain).
Again, negative knowledge is most important in the long run. Positive knowledge often gets replaced because it fails the tests imposed by new negative knowledge. Meanwhile, the negative knowledge persists : what counts as error changes much less often than what avoids error. When there are changes in negative knowledge, they tend to add to it rather than overturn old conceptions of error. Negative knowledge is the skeleton around which positive knowledge is organized and the base on which further development occurs.
**From our standpoint, expertise largely consists in a vast knowledge of sorts of errors that can occur and that can be made in myriads of specific circumstances within the domain. Non-experts will tend to blunder into such specific, perhaps rare, errors. All too often, that is how experts acquired their knowledge as well. If we want to understand how people become experts, we need to keep negative knowledge and its development at the forefront. **
4.1.1. Negative knowledge in training. Most forms of expertise require extensive teaching or apprenticeship, although neither guarantees attainment of the highest levels. Today negative knowledge rates little attention from teachers or trainers. Positive knowledge gets the spotlight, as might be expected in a culture like ours that is permeated with empiricism.
Such attitudes notwithstanding, teaching about misconceptions by presentingcountervailing examples and principles can be most useful. For instance, students’ misconceptions in science are a prevalent, persistent, and troubling phenomenon in education (Berliner and Casanova 1987, Brumby 1984, Champagne et al. 1983, Confrey 1990, Novak 1987, Perkins and Simmons 1988). Such misconceptions frequently maintain their hold on students, even after they have thoroughly mastered the standard positive-knowledge curriculum. Explicitly teaching negative knowledge helps students to overthrow such misconceptions
Moreover, we would argue that the errors human beings make in mastering a domain are not all avoidable imperfections, slips, or `performance limitations ’ . Instead, the best possible developmental pathway towards expertise necessarily incorporates errors.