Innovation is a key concept to both the behavioral and the evolutionary sciences for much the same reasons, yet the two fields differ in their definitions. Behavioral scientists do not currently distinguish behavioral “innovation” from “improvisation” or “invention,” identifying all of these as a novel behavior's first occurrence. By contrast, an evolutionary innovation is the spread and fixation of a developmental novelty throughout the population or group. Ramsey et al. make a helpful distinction in passing, between behavioral novelties and their cultural entrenchment (sect. 2). Yet they do not examine the dichotomy any further because they are primarily interested in invention, or as they call it, the “individual-level process definition of innovation.”
Whereas many researchers of behavior assign innovations to individuals (see Reader & Laland Reference Reader and Laland2003b), interdisciplinary biologists seek a consensus with the long-standing use of the term in evolutionary biology (Erwin & Krakauer Reference Erwin and Krakauer2004). The evolutionary meaning of innovation is applicable because behaviors are like other phenotypes, differing only in the larger role the environment often plays in their expression. Learned behavioral novelties and inventions are developmental novelties, which spread by environmental (rather than genetic) mechanisms such as public information (Danchin et al. Reference Danchin, Giraldeau, Valone and Wagner2004), social learning, or any other socio-ecological change, whether shared by mass migration or a shift in the environment.
The substantive contribution of Ramsey et al.'s article is its emphasis on the importance of individual learning and intelligence to behavioral innovations, as compared with social learning. It was once assumed that social learning improves fitness by exempting individuals from the costs of individual (asocial) learning. Individual learning carries high costs because trial and error, for example, can have fatal consequences in circumstances such as learning to forage and avoiding poisonous foods, or learning to fly and avoiding falling. However, Rogers (Reference Rogers1988) shows that sparing the costs of individual learning would not lead to increased fitness in the population. Like cheaters exploiting a social system based on the expectation of “altruistic” reciprocity, social learners increase their fitness when they are rare, but decrease overall fitness when they are common. Laland and Kendal (Reference Laland, Kendal, Fragaszy and Perry2003) describe how social learning continues to depend on individual learning even when conditions are most favorable for learning socially. Behavioral innovations depend on individual learning (e.g., conditioning and other forms of environmental induction) because social learning alone cannot track changes in the environment without individually learned input (Henrich & McElreath Reference Henrich and McElreath2003). But are these opposing strategies heritable, or is social learning just individual learning in the context of the social environment? Developmental psychologists working with primates, Fragaszy and Visalberghi (Reference Fragaszy and Visalberghi2001) found that the various kinds of learning are not specialized for exploiting particular kinds of information, and that “[t]here is no way to distinguish ‘social learning’ from ‘asocial’ learning … Socially biased learning is distinguished by the context in which learning occurs, not by distinctive cognitive processes” (p. 83).
If learning is a primary contribution to the generation of novelties that make successful evolutionary innovations, this calls into question the existence of separate vocabularies and explanatory methods and the opportunistic borrowing of concepts among “sibling” natural and behavioral sciences. Like biological mutations, novel behaviors can open the possibility both for their own eventual innovation and for effecting innovation in other traits. Behavior's effects include evolutionary patterns related to the Baldwin effect (Reference Baldwin1896a) and genetic assimilation (Waddington Reference Waddington1957), which at the limit may be irreversible, either because they constitute a “selective ratchet” preserving functional integration or an “epigenetic ratchet” in which the ancestral state is unregainable on account of feature loss or specialization (Futuyma Reference Futuyma1998, p. 693). Behavior can reinforce and even alter evolutionary trends through the building, maintenance, or shift of ecological niche (Odling-Smee et al. Reference Odling-Smee, Laland and Feldman2003). The capacity of the individual organism to affect evolution is a variable that is never at the level of chance, since organisms range from seeking adequacy (e.g., along chemical gradients) to intentionally optimizing their environments through invention. In particular, the choice of mates in sexually reproducing species can become a powerful force through the positive feedback of sexual selection, and have a runaway effect. Finally, behavior may affect evolution in its degree of plasticity or adaptability. Two examples are mobility, which allows for encounters with new possibilities for and conditions on evolutionary change (Lloyd Morgan Reference Lloyd Morgan1896; see Bateson Reference Bateson, Ho and Fox1988; Reference Bateson2004), and cognitive plasticity, which can be beneficial particularly in environments that allow trial and error. Contextualizing the study of behavior within the scope of evolutionary biology offers powerful tools in the discussion of innovation.
