

I’d like to see them take variance (i.e., range of different sentence lengths) into account—my impression is that more-modern writing has a more dynamic mix of long and short sentences, which could skew the stats in an unintuitive way.
Consider two paragraphs that both have a total of 300 words:
-
Paragraph A has three 100-word sentences.
-
Paragraph B has two 140-word sentences and four 5-word sentences.
Paragraph B has half the average sentence length of paragraph A, but over 90% of the text of B is comprised of sentences that are significantly longer than any of the sentences of A.
Accept that there’s going to be political diversity and social change (for better and worse) in spite of anything I might impose, and instead try to create an overarching framework to channel it into something other than violent conflict. One idea:
Let societies do whatever they want, but institute a “risk mitigation” tax (or other form of resource redistribution) based on size and similarity: if a social strategy is popular and widely adopted, it’s taxed at a marginally increasing rate until it reaches an equilibrium level; and the revenue is used to fund other, more experimental social strategies. This flips the historical dynamic on its head: instead of each society trying to forcibly convert the rest of the world to its own system, each society has an interest in discouraging others from following its example and trying something new instead.