Gould borrowed ‘spandrel’ from the architects, who said he had it wrong
The tapering triangles under the dome of San Marco gave biology its word for a trait that selection never built.
In architecture, the roughly triangular space between the outward curve of an arch and the squared frame around it, sometimes filled with ornament.1
A word lifted from the dome of San Marco
In September 1979, in the Proceedings of the Royal Society, Stephen
Jay Gould and Richard Lewontin opened an argument about evolution by
looking up at a ceiling.2
Both worked at Harvard’s Museum of Comparative Zoology, and both
were impatient with a habit they saw across their field: explaining
every trait of an organism as something natural selection had built
for a purpose. To make the point they went to the basilica of San
Marco in Venice. Mount a dome on four rounded arches and you are
forced to leave four tapering triangular spaces where the arches meet.
The mosaics fitted into those spaces are so elaborate that a visitor
might read them as the reason the spaces are there. That, Gould and
Lewontin wrote, inverts the analysis.
Spandrels do not exist to house the
evangelists.
2 The
space comes first, forced by the architecture, and the decoration only
uses it. In biology the word came to mark the same thing the ceiling
shows: a trait that is a necessary byproduct of some other feature,
not an adaptation built for a use of its own.2
The architects’ correction
A spandrel, strictly, is the flat space beside an arch.1 The curved triangles that actually carry a dome, the ones at San Marco, are something else: pendentives, segments of a sphere that turn a square bay into a round base for the dome to sit on.3 Sticklers for the technical vocabulary said as much, pointing out that Gould and Lewontin should have written pendentives.4 Gould answered them in 1997, and he began by conceding the building. He called the structures the pendentives of San Marco himself.5 His defense was about the word alone: it could stretch from the narrow architectural sense to a broader one, wide enough to cover the pendentives too. Biology, he argued, needed a single term for a feature that arises as a byproduct of something else, and the word was worth widening to get it.5
The term spandrel may be extended from its particular architectural use for two-dimensional byproducts to the generality of ‘spaces left over,’ a definition that properly includes the San Marco pendentives.5 Stephen Jay Gould, PNAS, 1997
Byproduct first, use later
A spandrel is the byproduct itself. When such a byproduct is later put to work, that use has a separate name: exaptation.5 A trait can start as a spandrel and become an exaptation once something co-opts it, but the two words fix different moments, and Gould kept them apart even while defending the first. The San Marco triangles are the byproduct; the mosaics are what later filled them. The everyday shorthand for a spandrel, a feature that came along for free and only later found a use, quietly runs the two together.
The word outgrows biology
The reason to keep the word is that it travels. By 1990 it had crossed into the study of language. In Behavioral and Brain Sciences, Steven Pinker and Paul Bloom gave a section the title “Arguments for Language Being a Spandrel.” They argued against the claim, credited to Gould, Chomsky, and others, that human language is a byproduct of a larger brain rather than a system selection shaped.6 They handled the word as a plain count noun, weighing whether language was one. The borrowing had done its work: a term coined for a ceiling in Venice now let Pinker and Bloom put a question about the origin of the human mind into a single word.6 That portability is why spandrel earns a place in a working vocabulary far outside the field that first reached for it.
Sources
- Merriam-Webster · spandrel
- Proceedings of the Royal Society B · The spandrels of San Marco and the Panglossian paradigm
- Encyclopædia Britannica · Pendentive
- Philip Steadman · The spandrels of San Marco, or rather the pendentives
- PNAS · The exaptive excellence of spandrels as a term and prototype
- Behavioral and Brain Sciences · Natural language and natural selection