Culture Is High Dimensional

Systems allow for more complex analysis, further from direct experience, using longer chains of reasoning. Such systems and their applications can be improved, and can more clearly be shown to have improved. And systems support finer and more complex divisions of intellectual labor.

The world of thinkers can be split into those who rely more versus less on established systems. Those who rely more on systems can be more precise, numerical, agree more on claims, and better evaluate each others’ abilities. Systems people less need metaphor to understand each other, and less need alliances, prestige, and social skills to coordinate with each other. (More)

[Before states,] people used concepts, norms, names, physical units, locations, and assets that were deeply entwined with local culture and practice. Concepts that states found hard to understand and apply, when they tried to enforce laws or extract taxes. So states, once they arose a few centuries ago, pushed people to instead use concepts that states could better see and apply from their bureaucratic distance. Such as unique names, global locations, and standardized languages, units, measures, laws, and accounting procedures. …

STEM uses concepts and systems that allow very different and widely separated things to be compared and analyzed in similar and consistent ways, using a “view from nowhere.” Arts/humanities, in contrast tend to have a stronger grip on our aesthetic, emotional, and moral reactions in particular situations and communities, when we very much do and want to see our world differently than do outsiders. (More)

The above are my prior attempts to make sense of the STEM vs humanities/arts distinction. Lately I’ve been pondering why so much talk on specific cultures (not cultural evolution in general) is so different from STEM talk. Such culture talk is said to be: symbolic, normative, qualitative, provocative, ambiguous, diffusive, resonant, synthetic, and interpretative. It uses analogies, metaphors, aphorisms, and “thick descriptions”. It more critiques, “understands from the inside”, and seeks new perspectives. To a STEM person, this is all “what the hell?!”

Let me try again, now analogizing to foraging in high dimensions:

Material of this universe, shaped into those floppy networks, has a complex hard-to-predict fractal geography … Because of the low density, high dimension, and unpredictability of motion, such creatures can’t use light to usefully see much besides the material they are at, or to communicate across material gaps. But they … can collect and share maps of the topology of the material connections between the places they have been. … high dimensionality allows ropes connecting creatures to be quite short and cheap, even when they cross enormous volumes of space and material. … an analogy to a species or civilization exploring a high dimensional space of possible cultures. (More)

In low dimensional topic spaces, it is more possible to describe stuff in terms of fewer standard descriptors. Which then allows one to collect datasets of comparable stuff, and to draw graphs and maps of it all in terms of such descriptors. And then to work out systems that usefully organize stuff, with standard ways to evaluate proposals to change such systems.

But in high dimensional spaces one often can’t find such standard descriptors. One instead has specific items connected in a network to specific neighboring items, and maybe if one is lucky one can find a limited set of standard connectors between items. One can take nearby sets of items and try to describe key ways in which they are the same or different, and then try to project those distinctions farther out into the network. But it is hard to do that in robust or stable ways. New connections project off into new dimensions, and all the configurations are floppy and changing.

When one finds different concepts that make sense of different but adjacent clusters, one often finds to relate those concepts, using candidate more general concepts. But there is often a large space of somewhat different versions that could work similarly well to unify those examples, but which would have quite different implications if projected further out into the network.

If one adds the habit of adding real connections between items, a vastly larger set of stuff can become “adjacent” to any one item. And then the resulting concepts that organize these purposely-connected items are closely connected to the reasons people used to add such connections. So then a few key concepts might organize a larger space of stuff, but only because those concepts directed the construction. There would still remain a vast larger space of possible things one could build, if only one would consider them.

Thus talk about stuff in large dimensional spaces uses thick descriptions of particular cases, especially the cases where people actually are, analogies and metaphors to compare neighboring clusters, and ambiguous concepts that have some prototypical examples, but which become less well pinned down as one moves away from those. It makes sense for people to spend a lot of time critiquing the concept proposals of others, and trying to interpret unusually successful concepts, to see what they could mean.

Of course it is quite possible that we will in the future find better more robust concepts that make STEM talk more feasible for lager areas of culture space. And this seems well worth the attempt. But until we do, I feel now more sympathetic to existing culture talk, as a way to deal with high dimensional spaces of culture topics.