What processes underlie the emergent complexity of human culture and cognition? Predominant theories of cultural evolution center bottom-up processes and marginalize the role of top-down design, suggesting that cultural complexity emerges from the accumulation of small, local innovations in response to immediate problems, rather than from any deliberate designer’s “grand plan” (e.g., Boyd & Richerson, Reference Boyd and Richerson1985; Blackmore, Reference Blackmore1999; Dawkins, Reference Dawkins1976; Henrich, Reference Henrich2016). Chater and Christiansen’s (C&C) theory of social tinkering offers an insightful, updated version of this view, but it remains a distinctly bottom-up theory. In the modern day, however, social tinkering and other forms of social learning often occur in designed environments, like business organizations and online platforms, which are deliberately engineered to elicit some behaviors and suppress others. Without accounting for the interaction between human cognition and designed environments, and the feedback loops and co-evolution between bottom-up and top-down processes, theories of cultural evolution may be incomplete.

Consider, for example, an organizational designer who specifies the network structure over which staff can communicate (e.g., an employee can communicate to their peers and line manager, but not directly to the chief executive officer) and the primary objectives of different teams and divisions (e.g., the sales division should optimize for revenue generation, while the customer-support division should optimize for service quality). Or, consider a social media platform engineer who specifies the rules of algorithmic amplification (e.g., the news feed algorithm might recommend content only from within a user’s network of “friends,” or it might recommend content from outside of the user’s network) and which kinds of communicative signals are allowed (e.g., users might be allowed to “like” others’ posts, but not “dislike” others’ posts). In neither case have properties of the environment been determined by selective cultural evolution over generations – they have been determined by human designers who purposely aim to steer others’ behavior from the top down, be it to turn a profit, maximize user engagement, or otherwise. Designers may be influenced by the bottom-up behaviors they observe and participate in, but an asymmetry exists: design is global, not local. Today, commercial and public policy designers can reshape choice environments almost overnight, with consequences for entire populations. In our view, reducing design decisions to merely another form of social tinkering fails to do justice to the power of choice environments and the myriad motivations that drive it. There may, of course, be downstream consequences of design decisions that designers themselves fail to anticipate, particularly as choice environments and behavior coevolve through feedback loops. But, the accuracy of the designers’ foresight does not change that they are designers manipulating the environment from the top down and thus influencing whatever bottom-up social learning occurs in those environments.

Indeed, the premise that top-down design can scaffold and steer bottom-up self-organization is central to a large body of existing literature on collective behavior and computer-supported cooperative work, which examines how the design of social technologies can elicit different emergent outcomes (e.g., Grudin, Reference Grudin1994; Goldstone & Gureckis, Reference Goldstone and Gureckis2009; Gregg, Reference Gregg2010; Malone, Laubacher, & Dellarocas, Reference Malone, Laubacher and Dellarocas2010; McKee et al., Reference McKee, Tacchetti, Bakker, Balaguer, Campbell-Gillingham, Everett and Botvinick2023; Smaldino et al., Reference Smaldino, Russell, Zefferman, Donath, Foster, Guilbeault, Hilbert, Hobson, Lerman, Miton, Moser, Lasser, Schmer-Galunder, Shapiro, Zhong and Patt2025). There, it is widely understood that although intelligent outcomes can emerge from collectives through “spontaneous order” in the absence of an explicit leader, it is not guaranteed. Rather, collective behavior is understood to be fragile and require deliberate stewardship (Bak-Coleman et al., Reference Bak-Coleman, Alfano, Barfuss, Bergstrom, Centeno, Couzin, Donges, Galesic, Gersick, Jacquet, Kao, Moran, Romanczuk, Rubenstein, Tombak, Van Bavel and Weber2021). For example, Salganik, Dodds, and Watts (Reference Salganik, Dodds and Watts2006) showed that collective cultural preferences for music emerged bottom-up through social learning among thousands of participants in an experimental marketplace but became increasingly arbitrary as designers amplified the visibility of others’ choices. That is, the same population converged on different “hits” depending on a parameter set by the experimenter (ibid.). Analogously, designers of real-world social media platforms implement global changes from the top-down that impact the local interactions of millions of individuals (e.g., Su, Sharma, & Goel, Reference Su, Sharma and Goel2016).

Recognizing the role of designed environments in cultural evolution has two key implications for C&C’s theory of social tinkering. First, it complicates the social tinkering view of cultural change as primarily bottom-up and gradual, because the conditions under which bottom-up tinkering occurs are often scaffolded by environmental properties, constraints, and affordances that are the product of top-down design. Second, it challenges the social tinkering view of individual cognition, which suggests that “the goal is to align not with an external reality but to coordinate with other people” (p. 10). It is not that people must coordinate with other people instead of aligning with an external reality, but rather that people must coordinate with other people and align with and navigate designed structures in their environment. This also is suggested by research on ecological rationality, which explains that the success of cognitive processes is determined by their degree of fit to the structure of the environment in which they are applied (e.g., Hertwig, Hoffrage, & the ABC Research Group, Reference Hertwig and Hoffrage2013; Hertwig et al., Reference Hertwig, Leuker, Pachur, Spiliopoulos and Pleskac2022; Todd & Gigerenzer, Reference Todd and Gigerenzer2007). For example, people’s tendency to infer that higher payoffs signal lower probabilities is ecologically rational in natural environments where this relationship regularly holds, but it leads to predictable losses in casinos and lotteries that are deliberately designed to exploit it (Pleskac & Hertwig, Reference Pleskac and Hertwig2014). Similarly, the success of social heuristics used by individuals to coordinate with one another is likely to be environment-dependent, meaning that it would be incomplete to explain human cognition purely in terms of coordination among individuals, independent of the external environment in which that coordination occurs.

In sum, C&C’s theory of social tinkering offers a valuable, updated theory of how coordinated, directed innovation among individuals can generate cultural complexity without deliberate planning. But, by treating cultural evolution as primarily bottom-up, as done by seminal theories of cultural evolution before it, C&C’s theory overlooks the extent to which modern commercial and public environments in which tinkering occurs have themselves been designed. A more complete account of cultural complexity must therefore integrate the bottom-up processes that C&C describe and the top-down design choices that scaffold and constrain them, as well as co-evolutionary dynamics.