Feedback for Rich on BB&A Research meeti

Feedback for Rich on BB&A Research meetids Research program 1: Sonic feedback

I have a good friend who’s a professor of math; he has spent a lot of time thinking about how mathematical ideas are embedded in sensory fields. I remember talking with him about someone who studies high-dimensional spaces, and whose “gears of my childhood” are the fields of interference between multiple audio speakers. For more points of contact with theories, I think Marshall McLuhann is headed in a similar direction when he talks about how we make sense through media, and how different media are essentially distinguished from one another by drawing on different “ratios of the senses.” And I think Finnegans Wake is all about this; there may be quite a bit of literary theory that you could draw on in framing experiments. 

I think this is an interesting learning sciences problem because it crosses the boundary from objective to phenomenological. I think it would be unsatisfying if your experimental setup didn’t engage with the phenomenological, if it just said people got more questions right in this setup than that setup. But then there’s the question of how we model the phenomenological, the shape of one’s ideas from one’s own perspective. This is very close to constructivism though. 

I think it would be a good idea to do some pilot studies in person before running it online; there may be embodied and situated factors that matter a lot. 

Research Program 2: Designing Virtual Experiences for Sense-Making and Inquiry

I think you should get us all invited to Bret Victor’s lab to explore this idea (https://harc.ycr.org/member/; http://limn.it/utopian-hacks/). Can we go in January?

I think the next step here is to decide which precedents you want to ground yourself in, and in which direction you want to make your inquiry. Is this about extending the simulation-for-learning research out of Northwestern and adding a haptic element? Is this about debugging? Is this based in science inquiry learning, adding the computational model? Or is this sort of a formalization of sociocultural practices already happening in Minecraft?

Among other things, this decision will shape what direction you go in designing the propositional language with which you model the world. We know that things quickly become intractable when you try to model the world (SHRDLU), so you’ll need to decide what to focus on. It seems like one possibility is to go the direction of Minecraft, and have an intentional/artistic lo-fi approach: big voxels, discrete number of elements and sorts of interactions possible between elements. 

Any chance you’re considering an AR approach? Going back to your first research program, we have such tremendous dependence on our embodied sensibilities that I think it’s exciting to use everyday real-life (and the social possibilities built-in) as your context. If you were considering this approach, one interesting possibility would be to support only a few model environments. Probably the first would be a flat plane, then maybe a flat plane going up against a wall, then maybe a staircase. Maybe you support nothing more than a parametrized ball (size, density, bounciness). Imagine if you could drop a ball five times, learn its parameters, and then replay it ad infinitum.