railways of the mind

hallucinations of agency

I’m still thinking about the relationship between cybernetics and politics, which is my big blind spot. I don’t think it’s a particular problem of mine – pretty much every structuralist theory since Karl Marx’s dialectical materialism has sooner or later run into a wall when someone asks “but seriously, is there absolutely no place for agency in this?”. But nonetheless, I feel obliged to have a go, even just to find out when and whether I need to give up and find a new obsession. And I think management cybernetics isn’t necessarily vulnerable to all the same criticisms as materialism, because information really is different.

Think about it this way; imagine that you’re opening up the American West in the 19^th century, and your big technology is railroads. If you’re a railroad baron, you have a lot of agency – you can decide where to lay the track, and that will shape the future of the continent.

But there are constraints! The physical facts about railroads and engines mean that it’s easier to put your tracks in some places than others. As well as politics, there are technical challenges; your choices have to take both kinds of constraint and tradeoff into account.

And you trade off priorities by making decisions; how much of your resources are going to be dedicated to overcoming particular technical and geographic challenges and how this interacts with the political environment you’re living in. To switch the context, think of the question of whether the route from London to Aberdeen should go via Dundee or not. The history of this is really interesting; it turns on a lot of politics and economics, the development of trade, and so on. But it also depends on the specific geography of the Tay and the Forth, and how they interact with technological progress in construction and bridge building.

So now switch context. Now you’re still a robber baron, but you’re trying to build part of the Ordinal Society. You’re building systems which collate and structure massive archives of personal data, in order to achieve some commercial or political goal, or some combination.

You have obvious political constraints on how you go about doing that. But there are also technical constraints! The algorithms that you choose, the way you store your data, the coding system and so on; these will make some kinds of structures and decisions easier than others. The slate heap effect will be present; architectural choices made early on when you don’t really know what you’re doing are quite likely to constrain future choices when you do.

The difference, I think, is that it’s much more difficult to make tradeoffs in this case than in the railroad one. Since the technological and architectural factors are to do with your information set itself, it’s much more difficult to get information about them. It’s much more difficult to analyse alternatives because it’s not just a matter of looking at a map and deciding how much you’re prepared to spend to go around a lake rather than building a bridge. The territory itself doesn’t exist until you’ve built it.

I think a consequence of this is that as systems get more complicated, and more of their energy and resources are taken up in matters of organisation and information, rather than physics, then agency is likely to become less important, as more and more decisions are constrained by factors which are hard to even recognise as constraints, because the capacity doesn’t exist to represent altenatives. Software eats the world, but it eventually eats the brain of the governance system.

Envoi: “We shape our tools, then our tools shape us” was a quote I considered as one of the chapter epigraphs for “The Unaccountability Machine”, but didn’t use in the end because it’s so bloody apocryphal