I was reading an aricle in Wired about machine translation and was struck by the connection to learning theory. The upshot was that new, much more accurate translation software is being written by creating context for the language being translated. The software takes a section of text in Spanish, say 8 words long, generates a list of possible translations in English and looks on the web for the frequency of each translations appearance in English as a measure of likelihood it is the best translation. It then moves one word ahead and repeats the process with seven words from the last set, plus the next one in the sequence. It continues like this selecting the most likely candidates for each set of 8 words and then produces a translation based on the most likely of all possibilities. Essentially what the algorithm does through brute force is create an understanding of language in context.

The reason that I think this is so interesting is that it supports the idea language is contextual. As language is the primary tool for learning, by extention this means that learning (or knowledge) is contextual as well. If you can’t learn the word “bank” without having a context for the word, then you can’t learn about sedimentation collecting on the bank of a stream without that taking place in a context. The context of the learning is as critical as the content. For me this shatters the idea that is fundamental to so much of how we assess learning (e.g. NCLB & standardized tests) — that knowledge out of context has meaning and that measuring knowledge out of context is a meaningful way of understanding what a student knows. If a computer cannot understand things well without a context, must it not be even more true of people?

I have been reading the new Cambridge Handbook of the Learning Sciences. In his introduction the editor, Keith Sawyer, mentions untested ideas about learning. One of the ones that I found most critical is the idea that the sequence and content of ideas children should learn is decided by (content) experts, not by studying how children actually learn. What this means is that high school Physics, for example, is taught based on how experts in the field of Physics see the organization of Physics as a discipline. This in no way represents any consideration of how these concepts are best learned. In fact it does not even represent how human beings, over the course of the development of Physics, have developed these ideas. So, for example, would it makes sense to start teaching kids Physics by starting with the ideas the Greeks had about it and then showing them how those ideas were overturned? Maybe or maybe not. The point is we are not basing our choices on how students learn best, but on how experts view their field. Who then should be making decisions about how Physics should be taught? Well (no surprise here), people like me. Those of us who instead of studying Physics study how people learn Physics and thus how to best teach it.

This highlights one of the most difficult things about expertise in education – it is widely considered to be part of everyones everyday experience, and thus true “expertise” in teaching and learning is suspect. Physics professors assume they know something about education because they are assigned to teach a section of introductory Physics. How do they usually choose to teach? The way they were taught. If you did this in Physics — do things just like your teacher had done them, without ever testing it empirically — we would still believe what Aristotle did about the world (or the cavemen). It simply does not makes sense as a way to improve our understandings of the world.

The counterintuitive notion for me is that the discipline (e.g. Physics) has taken decades if not centuries to develop, and it has changed over time as naïve or intuitive notions about the world have been emirically tested. Most students, however, have many of the same naive notions about the world, but we assume giving them the current structure of the field will clear all that up for them. We need to understand how people learn and use that understanding to help them move from naive conceptions to deep conceptual understanding of the ideas in a discipline. This is not the way that we think about schooling.

In their column for the NYT magazine Freakonomics, Stephen J. Dubner and Steven D. Levitt discussed the problem of getting doctors to wash their hands in the hospital. For me there are clear parallels between this phenomenon and NCLB (No Child Left Behind). Doctors were given free Purell hand sanitizer as they pulled into the parking lot, they were “caught” being good (i.e. washing their hands before examining a patient) and given $10 Starbucks giftcards. All this led to an increase from 65% to 80% proficiency in hand washing. Finally, they cultured the doctors hands and then made a screen saver for every computer in the hospital with a picture of the massive bacterial colonies that resulted. This brought proficiency up to 100% (they claim) and it has stayed at that level since. Dubner and Levitt then point out how the solution to a seemingly simple problem (getting doctors to wash their hand before examining a patient) is often incredibly difficult and time consuming.

So, what has this to do with education and NCLB? Well, for me it brings out the massive underestimation of the complexity of learning and teaching on the part of almost everyone. This is exemplified in NCLB, which asks for 100% proficiency among students and proposes punitive measures based on standardized test scores as the motivation. If we compare the complexity of getting all american school children to 100% proficiency to getting doctors to wash their hands, it seems pretty obvious the differences in complexity. Yet, with the doctors all the cohersion (both rewards and punishments) could only get them to 80%. Do we really think that we can legislate a solution to the problem of schools failing to prepare our children? If there are kindergarten classrooms with 35+ students taught by someone with emergency certification (i.e. no preparation to teach), what is a realistic level of proficiency for those students? I don’t know what the equivilent of the dirty hand as screen saver is for our educational system, but NCLB is not it.