Languages are cultural traditions that combine great complexity with universal emergence across diverse environments. Singh defines super-attractors as sociocultural phenomena representing packages of functionally interrelated attractors. Languages fit this characterization precisely. They are packages of words and constructions, each itself an attractor maintained through cultural transmission. The word “cat,” for instance, is an attractor comprising token sounds similar to /kat/ associated with mental representations similar to [CAT]. Languages emerge reliably in every known culture. In fact, language may be the paradigmatic super-attractor even though Singh makes no mention of it.

Can language evolution be explained by subjective selection and subjective selection alone? Subjective selection refers to the production and selective retention of cultural variants evaluated as instrumentally useful for satisfying goals. Studying subjective selection means privileging explanations that treat agents as goal-directed and endowed with psychological mechanisms making them effective at achieving these goals. We find this approach promising and useful because it applies methodological individualism to explain social phenomena and also their evolution over time.

Languages are culturally evolved technologies for satisfying speakers’ informative intentions. Singh characterizes cultural traditions whose function is achieving goals as technologies. This works well for language: speakers craft utterances to achieve communicative goals, primarily to affect their audience’s mind by conveying intended meanings. Thus, many linguistic features indeed result from subjective selection.

The terms a language develops are precisely terms about things people want to discuss – terms helping them achieve informative goals. Kinship terminology provides a compelling example. Rather than simply tracking genetic relatedness, kinship systems reflect social relationships that matter for predicting behavior, cooperation, and resource allocation. People develop kinship terms representing information about genealogical distance, social rank, gender, and group membership – precisely the information having predictive validity for understanding how individuals will behave toward one another. This likely represents subjective selection as speakers retain linguistic distinctions serving their instrumental goals to shape and refer to social relations (Jones, Reference Jones2003; Nettle & Heintz, Reference Nettle and Heintz2025).

Differential case marking provides another example: it refers to the fact that nouns are marked for their grammatical role only if this information cannot be derived from world knowledge. This also likely results from speakers choosing to minimize effort while maximizing communicative success. When ambiguity threatens communication, speakers add marking; when meaning is recoverable from context, they omit it. Smith and Culbertson (Reference Smith and Culbertson2025) demonstrate that these efficient patterns emerge specifically during communicative interaction (rather than at the learning stage), supporting a subjective selection account.

However, subjective selection has limits. Singh’s framework privileges conscious, goal-directed behavior, so its explanations are limited to behavior resulting from individual motivations. This leaves out the many unintended effects of behavior, which contribute to shaping cultural phenomena. Additionally, intentions and motivations themselves need explanation. Understanding why people have the goals they have contributes to explaining cultural evolution and falls outside subjective selection’s explanatory reach. It falls in the domain of evolutionary psychology, and thus, subjective selection must be coupled with accounts of the evolved architecture of the mind.

Sperber’s (Reference Sperber, Boudon, Chazel and Bouvier1997) analysis of the factors that shape cultural evolution emphasize the necessity to include “sub-personal” factors: mental processes not always reducible to belief-desire psychology; they include mental capacities and their properties, whether related to learning, motivation, memory, or attention. Cultural evolution studies require attention to cognitive processes operating below conscious intention, leading social scientists to better integrate findings from cognitive psychology. Sperber thus advocates for “infra-individualism” as going beyond methodological individualism.

Vowel spacing provides an example for infra-individualism. Languages tend to preserve maximal phonetic dispersion in their vowels: that is, the centroids of the distinction vowels are as distant from one another as possible given the number of vowels and available acoustic space (Liljencrants & Linblom, Reference Liljencrants and Lindblom1972). Yet, no speaker decides to have nicely spaced vowels. Rather, the regularity is a consequence of a low-level mechanism called peak shift. Tokens of a vowel X that fall closer to the centroid of vowel Y are usually misclassified by learners as tokens of Y. Thus, the centroid of X that they learn is biased relative to the actual production distribution by other speakers, in the direction of being further from the centroid of Y (Labov, Reference Labov1994). The emergent effect of this in populations of learners is the gradual repulsion of vowels in acoustic space (or, their merger). Subjective goals play no direct role in the explanation. Learning biases shape language evolution in ways that cannot be reduced to speakers’ conscious intentions. Weak biases, which can operate below conscious awareness in individual learners, can nonetheless have strong effects on socially learned systems’ structure (Kirby, Griffiths, & Smith, Reference Kirby, Griffiths and Smith2014).

Finally, the term “subjective selection” may be misleading. It evokes natural or artificial selection, implying that random mutation creates options and choosers simply favor one over another. Processes of cultural evolution, however, involve agents creating and transforming variants in either goal-directed or unconscious ways, not merely selecting among random mutations. Consider the evolution of slang. New slang terms are often deliberately invented to achieve specific social goals – signaling group membership, establishing social identity, and preventing outgroups – the bourgeois – to understand what is said. The inventions involve computations of what kinds of substitutions will be funny, obscure, humiliating, apt, etc., and so the innovations are far from random.

A complete understanding of language as a super-attractor is a rich research field, which – under the name of “language evolution” research – has made many advances in recent years. The field recognizes that subjective selection plays a role, but also identified many factors that drive systematic transformation and non-intentional processes related to mental processes such as learning, remembering, and perceiving. Singh’s target article provides a useful reminder that analyzing human behavior as goal-directed has great explanatory power. Still, infra-individualism adds tools from cognitive psychology to the social scientist’s toolkit.