260714: Mountaineer in Fog
Perhaps we must be like the mountaineer, on a mountain range of fog, and seeking the highest peak.
Extensive simulations of such mountaineers have been run by mathematicians. As it turns out, one does not simply climb uphill, or else risks being trapped on the nearest lowly peak. Instead, we ought to wander early on—often walking in what appear to be downhill directions, or side-to-side. (It is not optimal to be optimal; it is not in our interest, to be greedy.)
And when, in our youthful explorations, we notice the morning sun approaching noon, and think of evening's dark? We change our tack, and stride uphill only—committed, never looking back.
This algorithm is known as simulated annealing. It is popular in the training of machine learning models. It is popular as a conceptual model for the human cognitive development from infancy through adulthood, and as a theory for the effects of aging, psychedelics, hormones, and post-traumatic stress.
If we try to apply it to the question of cultural change, we face three challenges. The first is nearly fatal. Elevation is a single, objective, measurable axis of comparison. Second, culture is an infinite game, which has no end, no falling dusk or death which might constrain its period of youthful wandering. Third—and this, at least, we can learn from—is the fact that real mountaineers identify and follow an elevated ridgeline, never dipping below it, and easing their survey.
Simulated annealing does provide one elegant parable. Far better than a single mountaineer are many hill-climbers, starting out from different positions, with different preferences, strategies, heuristics. After a period of climbing, the mountaineers compare their elevations.