I drive my oldest daughter to high school everyday. She goes to a magnet STEM school in the district that’s on the campus where I work. I’ve been brainwashing her into liking indie rock one car ride at a time using carefully planned mix CDs.

Last week she tells me I need to get more Magnetic Fields songs. Why? I ask.

“Physics homework.” she says.

It turns out that there’s a number of principles of physics that she remembers through a complex set of associations she’s developed referencing indie rock songs. I don’t pretend to get them all, but the 69 Love Songs hit “Meaningless” plays an apparently crucial role.

Later on my youngest daughter is asking me about the book Persepolis. The author of that book spends the preface talking about the reasons she wrote it, and how she felt the understanding of her native country of Iran was too narrow, and in a way, too exotic. She tells me that she doesn’t quite get what the author is talking about. After all, there’s a lot of fundamentalism in the early parts of the book — people are really in a revolution in 1978, so what are we getting wrong?

I know that this daughter, a middle schooler, has had some stress about Donald Trump. She has people in her class who like him, and she can’t understand why when he’s so mean. It worries her.

I ask her if Trump gets elected, how would she feel if everyone assumed all Americans were like Donald Trump. Well, we wouldn’t be, she says.

Oh, she says.

When we talk personalization, we tend to talk about targeting. You learn a certain set of things, you get tested, the personalization software finds knowledge gaps and runs you through the explanation that you need. (There are other, related meanings, of which there is a partial taxonomy here).

The idea seems to be that there is a wide variety in what concepts students struggle with, but there is one perfect explanation per concept. Personalization gets the explanation of that concept to the student.

That’s a part of the story, but it’s not even the most important half.

When tutors work with students they do alter what they work on based on student need. They personalize what skills they target, sure. But the big advantage of a tutor is not that they personalize the task, it’s that they personalize the explanation. They look into the eyes of the other person and try to understand what material the student has locked in their head that could be leveraged into new understandings.

If you find yourself teaching people something — anything — you’ll see this at work. How many times do you being with the phrase “So have you heard of X?” There you are, looking for the way into the explanation. It could be from point X, a Magnetic Fields song. Or from point Y, a Trump analogy. For a Trump-supporting indie-rock-hater it’s going to be a completely different entry point, and a different explanation.

This gets to my obsession with thinking about Open Educational Resources as explanations and data organized as variants around a namespace. Instead of us having curated and published the supposed “best” explanation of a subject, why not take a git-like approach, and let different explanations proliferate? Instead of “read this chapter on Broken Windows Theory” say “Here’s five variant explanations of Broken Windows Theory, and fifteen related examples — find the ones that work for you. and copy the ones that work best to your own space.”

Over time, what happens? People with similar backgrounds, needs, abilities, and talents as me curate their favorite explanations and examples to their space, and as I discover these people I discover a hand-crafted curriculum for me, one that makes use of the fact I have an encyclopedic knowledge of indie music and a background in political theory to teach me new things. The resources you tap into teach you the same things as me, but from different starting points.

Accessibility is baked in, as some resources are particularly good for students with visual impairment. Some might even use a students unique experience of the world as a strength. Can you imagine an explanation for a deaf student that doesn’t just work around the disability, but occasionally says something like “You know how when you’re deaf..”

Adult learners might get examples and writings that don’t treat them as a 19 old. Business majors might get explanations of psychological concepts that apply to business. These modules would be recombinable and remixable into unlimited combinations, and each explanation would be linked to its variants.

This is really the power of OER that has not yet been unleashed.  The problem, I think, has been getting people who haven’t used federated wiki or Wikity to understand what a system of connected copies like this might look like. Thoughts?