A thought experiment on thought experiments

The limitations of argument by analogy, particularly of historical analogy, are well-known. Strands of apparent likeness between two situations may be far outstripped by myriad factors that are in no way like.

But of course, we can't do without analogy. All inductive reasoning is dependent on it. Past results may be no guarantee of future success. But everything we know about the course of human events derives from knowledge of the past.

More broadly, every noun and every adjective is an analogy, or rather a classification, which is an analogy extended across a number of discrete entities. A noun is an assertion that different entities are like enough to share a name; an adjective, that different instances of a quality (brown, angry) are like enough to be classed together. For that matter, I suppose a verb embeds an analogy too -- the rapid motion of your legs is like enough to mine that we can both be said to "run," even if you move twice as fast as I do.


And ah, the eureka of a really good analogy! What comes first to mind (through nether pathways) is the instructive assertion that giving birth to a human baby is like "shitting a watermelon." Less viscerally, what I'm meandering toward  is mentally yoking two really excellent extended analogies that I've come across recently -- each complex enough to be deemed a "thought experiment," as I suppose any really good analogy is. If I can find any common element other than aptness (I'll think about it as I paste them, which doesn't work as well for this purpose as typing them out), I suppose I'll have my own meta-analogy.

First: James Fallows recently relayed a sterling attempt by consultant Thomas Barnett to help Americans imagine the strengths and weaknesses of a rising China: