The Infrastructure of Intention

So I am starting a new and rather different job tomorrow. In preparation for that, I have been ruminating on an topic for the last few weeks, trying to turn it into a post before I start and get immersed in a new and probably overwhelming environment.

This idea is nicely captured by the phrase "The infrastructure of intention". That is to say, living things, social systems, technologies, all embody purpose in various ways and all these purposes have various ways of interacting with each other and wouldn't it be great if we could find some better ways of both analyzing these extremely important processes, and improving them? I didn't get very far with this essay because the ideas are just way too big for a blog post, and kept threatening to grow into something dissertation-sized.

Fortunately in goofing off from addressing it I ran into this post by Robin Hanson, which very helpfully reminded me that questions are typically more interesting than answers. So, here are some questions around the idea of intention and computation, some of which have been dogging me for decades. Some managed to get into my actual dissertation, and some others may be addressed in this new gig, but we will see. I'm just a computer programmer, which means most of what I do is just informational plumbing, and it doesn't leave that much time for grandiose theorizing. But high and low have their ways of coming together on occasion.

So, the questions (and pointers to people who have spent more time thinking about them then I have):

• What is the nature of purpose? (Cybernetics, particularly Gregory Bateson)

• How do humans (and animals manage their various divergent intentions? (Freud, Tinbergen, Minsky)

• Can inanimate things have purpose? (Latour, Bennett)
• How do individual goals relate to social structures and institutions? (all of sociology and political science, at the moment particularly Charles Tilly and Mary Douglas)

• How do computer programs embody purpose? (AI)
• Can/should goals be explicitly represented? (Agre and Chapman's situated action critique of planning-based AI)

• How does can encode divergent and conflicting goals? (Actor theory, Multi-agent AI),

• How does goal-directed behavior work in human activity that is clearly non-functional in any simple way, like religion and art? (Evolutionary psychology)

• Can/should/how can software embody and extend human goal structures (the CSCW field, but originating maybe with Doug Engelbart)

• What would the world look like if computational infrastructure actually supported goals in a powerful way (lots of science fiction, mostly with a dystopian flavor, but for a somewhat more cheery spin, Bruce Sterling's story Maneki Neko)

Big fucking questions, aren't they? And quite out of scale compared with my ability to provide answers, but they won't leave me alone.