Comments on “Diverse forms of agency”

I think we can probably usefully abstract agency as an emergent phenomenon. A guy I used to know invented the idea of an “emergent conspiracy” which is one in which all the conspirators are unwitting. These things are surprisingly common.

Institutional agency is, I think, pretty clearly modeled as something that is definitely agency, but which arises as an emergent phenomenon (see, for example, non-profits acting to perpetuate the problems that they are ostensibly intended to eliminate.)

There are models of cognition that suggest agency is just an emergent brain thing.

Anyways, when you get an interconnected set of things, at least some of which are loosely goal driven - call them optimizers - the thing as a whole can wind up as a complex machine for lurching toward a goal or collection of relate goals. The immune system is such a thing.

Salient features of these systems: open versus closed loop; friction

(I expect to run across these later on as I keep reading)

End of one paragraph and beginning of next have an awkward repetition:

“Institutions differ from the immune system in that some of their parts have their own beliefs, desires, and intentions.

Institutions also have their own beliefs, desires, and intentions, which may differ from those of their members.”

Do you want these kinds of comments? Let me know if this is unhelpful.

I have not explored that space extensively either, nor has anyone else as far as I know, so I can’t provide even a summary catalog.

Well, I’ve been obsessing about alternate forms of agency for years and have built up a kind of catalog of agency patterns: https://www.ribbonfarm.com/2012/11/27/patterns-of-refactored-agency/

Lots more at http://hyperphor.com/ammdi/agency

I wholeheartedly agree with everything on this page, except I would add: while it is certainly true that Nazism and communism are group agencies of a sort – so is capitalism. Capitalism is probably a more powerful and important form of agency, more successful, and certainly more relevant to understanding AI.

So, you’ve been galvanized into finishing a book. Mazel tov. I feel the frustration of years of being a starter and not a finisher, and haven’t broken out of it, nor even managed to keep something fascinating but unfinished for decades.

I’ve gotten through a first reading. I do think reading Shoshana Zuboff would add some depth to your ideas. She has also worked closely with Maria Ressa, a Philipina journalist with an extraordinary grasp of the dynamics of social media and she has used it in the struggle with the last few years of Philipine dictators. Probably Zuboff could convince you of this faster than I could.

You are welcome to “acephalous”, it’s not really my term, it’s basically anthropologist-speak for “anarchist”. Also an odd secret society formed by Georges Bataille https://en.wikipedia.org/wiki/Ac%C3%A9phale

“Egregore” has somewhat different connotations: of artificiality and constructedness. An egregore is distributed but its agency is thought of as singular, that’s what makes it a thing. Santa Clause is an egregore, and he is watching and making a list, despite his distributed implementation.

I think weirdness is unavoidable when discussing this stuff. None of these languages are quite right; agency remains a slippery and elusive concept.