Follow-up from Jonathan Wonham

Jonathan Wonham, whose post on Roger Caillois and the Mexican jumping bean inspired two posts from me last week (here and here), has posted a strong follow-up on Connaissances that I would recommend. Among the many interesting points in the post are Jonathan's comments on the distinction between (scientific) analogies and (poetic) metaphors:

So actually, I think most scientists would be suspicious of metaphors, especially those based on anthropomorphising 'feelings'. What they are often up for, however, is analogies. A good analogy is often used for explaining a complex phenomenon in terms of a simpler one. A good example in the domain of fluid flow is the use of the flow of traffic into a city as an analogy for the movement of a turbulent flow. I've only been able to find one reference to this analogy on the internet and it is buried somewhere here.

In this analogy, the cars on the road stand in for grains of sediment in the turbulent flow because the behaviour of the cars is much easier to examine than the grains of sediment. The scientist looks to see what happens to the speed of the cars as they approach the city, slowing down or speeding up as the traffic reaches various bottlenecks and then makes the analogy with the particles in the flow, suggesting that these will also slow down or speed up as the flow encounters similar 'bottlenecks' or constricting gullies on the sea floor.

Why do scientists prefer to call these comparisons analogies rather than metaphors? It is because, as made clear here, metaphor is a rather wide term which:

is not always used for practical description and understanding; sometimes it is used for purely aesthetic reasons.

Yes, but is the desire to avoid association with aestheticism enough to support a strong distinction between analogy and metaphor? Isn't the distinction to some extent semantic?

(There are many other points in Jonathan's post, including a very intelligent reading of Caillois's "Siliceous Concretions" poem as well as some disagreements with some of the arguments I've made. Read his whole post).