This is a classic paper criticizing the popular idea that evolution works by optimizing traits, proposing an adaptive story for each trait. Instead, the authors argue that organisms must be analyzed as integrated wholes with a history. They start with the example of spandrels in San Marco’s cathedral: there are paintings that perfectly fit those spandrels, as if the spandrels were part of the depicted story, as if they were designed to fit the story. But of course the reality is the spandrels are an architectural constraint, which has been exploited creatively by the painter. So the mere fact that some trait seems to be useful for an organism does not mean that it results from adaptation. Gould & Lewontin expose a number of flaws in adaptationist reasoning (depicted as story telling), then list alternative types of explanation.

This reminds me of an insightful remark by Mark Bickhard on evolution (Bickhard 2009). It is typical to say that some part of an organism, some organ, has a function in the sense that it has evolved to fulfill that function. But Bickhard points out that this is circular: it starts from the premises that normativity is preexistent (there is already a function) and therefore it cannot explain how it emerges (eg saying that the human vocal tract is adapted to language is weird, since language did not exist before there were speakers). It is better to talk of serving a function. For example the hand serves the function of typing on a keyboard, but it does not “have” the function of typing on a keyboard, as in something that is preexistent to the interaction with a keyboard. It just so happens that the hand can be used to type on a keyboard, and then possibly we might imagine that hands then evolve in a way that makes them better fitted to keyboard typing, but serving the function must be prior to having the function.

I wrote a related note on optimization as a metaphor of evolution.