Origin of the term
The term was coined by the Harvard paleontologist Stephen Jay Gould and population geneticist Richard Lewontin in their influential paper "The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme" (1979).
In their paper  named as "exaptations" those characteristics that enhance fitness in their present role but were not built for this role by natural selection. Exaptations may be divided into two subcategories; preadaptation and spandrels. Spandrels are characteristics that did not originate by the direct action of natural selection and that were later co-opted for a current use. Gould saw the term to be optimally suited for evolutionary biology for “the concept of a nonadaptive architectural by-product of definite and necessary form – a structure of particular size and shape that then becomes available for later and secondary utility” (Gould 1997)
Criticism of the term
Their suggestive proposal generated a large literature of critique, which Gould characterised (Gould 1997) as being grounded in two ways. First, a terminological claim was offered that the "spandrels" of Basilica di San Marco were not spandrels at all, but rather were pendentives. Gould (1997) responded, "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."
Other critics, such as Daniel Dennett, further claimed that these pendentives are not merely architectural by-products as Gould and Lewontin supposed. Dennett argues that alternatives to pendentives, such as corbels or squinches would have served equally well from an architectural standpoint, but pendentives were deliberately selected due to their aesthetic value. Critics argue that Lewontin and Gould's oversight in this regard illustrates their underestimation of the pervasiveness of adaptations found in nature.
Ian Kluge criticizes the whole subject of spandrels to be bogged down in a definitional debate, i.e. it is not entirely clear what is a spandrel and what isn't, and that, worse, all examples of spandrels, pendentives, corbels and squinches do actually serve a function, i.e. they are necessary to achieve something – but that necessity is exactly what epiphenomenalism denies.
Response to criticism
Gould responded that critics ignore that later selective value is a separate issue from origination as necessary consequences of structure; he summarised his use of the term 'spandrel' in 1997: "Evolutionary biology needs such an explicit term for features arising as byproducts, rather than adaptations, whatever their subsequent exaptive utility.... Causes of historical origin must always be separated from current utilities; their conflation has seriously hampered the evolutionary analysis of form in the history of life." Gould cites the masculinized genitalia of female hyenas and the brooding chamber of some snails as examples of evolutionary spandrels. (Gould 1997:Abstract)
Gould (1991) outlines some considerations for grounds for assigning or denying a structure the status of spandrel pointing first to the fact that a structure which originated as a spandrel through primary exaptation may have been further crafted for its current utility by a suite of secondary adaptations, thus the grounds of how well crafted a structure is for a function cannot be used as a ground for assigning or denying spandrel status. The nature of the current utility of a structure also does not provide a basis for assigning or denying spandrel status nor does he see the origin of a structure as having any relationship to the extent or vitality of a later co-opted role but places importance on the later evolutionary meaning of a structure. This seems to imply that the design and secondary utilization of spandrels may feed back into the evolutionary process and thus determine major features of the entire structure. The grounds Gould (1997, pp. 10752-10753) does accept to have validity in assigning or denying a structure the status of spandrel are historical order and comparative anatomy. Historical order involves the use of historical evidence to determine which feature arose as a primary adaptation and which one appeared subsequently as a co-opted by-product. In the absence of historical evidence, inferences are drawn about the evolution of a structure through comparative anatomy. Evidence is obtained by comparing current examples of the structure in a cladistic context and by subsequently trying to determine a historical order from the distribution yielded by tabulation.
Language as a spandrel
The linguist Noam Chomsky has argued that the 'language faculty', specifically the property of discrete infinity or recursion that plays a central role in his theory of Universal Grammar, may have evolved as a spandrel: in this view, Chomsky initially pointed to language being a result of increased brain size and increasing complexity, though he provides no definitive answers as to what factors led to the brain attaining the size and complexity of which discrete infinity is a consequence. Steven Pinker and Ray Jackendoff say Chomsky's case is 'unconvincing' and that 'language maps among recursive systems rather than being a straightforward externalization of a single recursive system', and as an example, numerical recursion ‘is parasitic on language (rather than vice-versa)’ among other arguments (Pinker and Jackendoff, 2005). Pinker contends that the language faculty is not a spandrel, but rather a result of natural selection (Pinker and Bloom, 1990). Newmeyer (1998) instead views the lack of symmetry, irregularity and idiosyncrasy that Universal Grammar tolerates and the widely different principles of organization of its various subcomponents and consequent wide variety of linking rules relating them as evidence that such design features do not qualify as an exaptation. He suggests that Universal Grammar cannot be derivative and autonomous at the same time and that Chomsky wants language to be an epiphenomenon and an ‘organ’ at the same time, where an organ is defined as a product of a dedicated genetic blueprint. Rudolph Botha counters that Chomsky has offered his conception of the feature of recursion but not a theory of the evolution of the language faculty as a whole. (Botha, 2001)
Similarly, Pinker has dismissed music as another spandrel to the enlarged brain.  Dunbar found this conclusion odd, and stated that "it falls foul of what we might refer to as the Spandrel Fallacy: 'I haven't really had time to determine empirically whether or not something has a function, so I'll conclude that it can't possibly have one.'"  He devotes the chapter to two potential roles of music in evolution: "One is its role in mating and mate choice, the other is its role in social bonding."
- Stephen Jay Gould and Richard C. Lewontin. "The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme" Proc. Roy. Soc. London B 205 (1979) pp. 581–598
- Stephen Jay Gould (1997). "The exaptive excellence of spandrels as a term and prototype" Proceedings of the National Academy of Sciences USA. 94: 10750-10755.
- Stephen Jay Gould (2002). The Structure of Evolutionary Theory. Belknap Press of Harvard University Press, Cambridge, Mass.
- Phillip Stevens Thurtle. "The G Files: Linking 'The Selfish Gene' And 'The Thinking Reed'"
- Daniel Dennett (1995). Darwin's Dangerous Idea: Evolution and the Meanings of Life. Simon & Schuster. ISBN 0-684-82471-X.
- Marc D. Hauser, Noam Chomsky, and W. Tecumseh Fitch (2002). "The Faculty of Language: What Is It, Who Has It, and How Did It Evolve?" Science 298:1569–1579.
- Robert Mark (1996). "Architecture and Evolution" American Scientist (July–August): 383–389.
- Steven Pinker, Ray Jackendoff. (2005) The Faculty of Language: What's Special About It? Cognition 95: 201–236
- Steven Pinker, Paul Bloom (psychologist). (1990) Natural Language and Natural Selection Behavioral and Brain Sciences 13(4): 707–784
- Rudolf P. Botha (2001). How much of language, if any, came about in the same sort of way as the brooding chamber in snails? Language & Communication 21:225–243.
- Stephen Jay Gould (1997). "The Exaptive Excellence of Spandrels as a Term and Prototype" (http://www.pnas.org/content/94/20/10750.full)
- Evolutionary biology uses the term Spandrel for features arising as byproducts, rather than adaptations, that have no clear advantage towards and organisms fitness or survival. Gould and Lewontin rebuted certain counter arguments that stated spandrels were just small unimportant byproducts but stated that, “we must not recognize that small means unimportant. Spandrels can be as prominent as primary adaptions” (Gould and Lewontin). A main example used by Gould and Lewontin is the example of the human brain. It is explained that the human brain is the area in humans that is thought to have the most spandrels. So many secondary processes and actions come in addition to the human brain and its main functions. These secondary processes and thoughts are the spandrels of the human body, which eventually through thought and action can turn into an adaption or fitness advantage to humans. Just because something is a secondary trait or byproduct of an adaption does not mean it has no use because it has the possibility to eventually be used as something beneficial towards the animal.
- S. Pinker, How the Mind Works. Harmondsworth: Penguin.
- Robin Dunbar, "On the Evolutionary Function of Song and Dance," in N. Bannan (ed.) Music, Language, and Human Evolution. Oxford: O.U.P., 2012, p. 201-02.