Moffett's goal of facilitating cross-disciplinary research by providing a definition of society that will translate across academic disciplines, as well as species, is both laudable and illuminating. However, the definition he provides has several fundamental problems based on an insufficient consideration of properties and processes of identity as explored by anthropologists and others over time, and so, I believe, will require refinement.

First, Moffett argues that societies should be thought of as identity groups rather than as groups based on shared culture (non-genetic information). However, as he discusses, people typically mark group identity via cultural markers – language, rituals, dress, food, and so on. Therefore, his concept of society is in many respects redundant, as simply using participants' perception of shared culture as the defining characteristic of (at least human) social groups will suffice. Moffett might object to this point because he notes that societies often “incorporate cultures from various sources,” but all social groups incorporate cultural traits from many sources, and even in the most homogenous contexts, individuals are differentially members of several cultural (or if you prefer, subcultural) groups simultaneously. What matters is which culture traits are selected to be used to signal group identity, irrespective of source or, for that matter, authenticity. Thus, for example, musical markers may be based on purportedly authentic history that is factually incorrect (e.g., Manuel, Reference Manuel1994), and traditions can be invented to serve identity-related purposes (e.g., Hanson, Reference Hanson1989). Markers are arbitrary, meaningful only to the extent they are agreed upon as legitimate, and people routinely debate their legitimacy (Qirko, Reference Qirko2014). And although they are most commonly thought of in the context of the social construction of ethnicity, race, and nationalism, markers are fundamental to any form of social group identity. Much of this Moffett discusses, but in a way that underestimates the degree to which identity markers and affiliations are contested, negotiated, revised, rejected, and replaced, both within and between groups, by parties with competing interests (e.g., De Fina, Reference De Fina, Locher and Graham2010; Nagel, Reference Nagel1994; Tajfel, Reference Tajfel1978). Instead, he argues that societies as he defines them have clear membership, comprised of people who see themselves as belonging together, where “no one is accepted as a fellow member by some and rejected by others.” But only under temporary conditions of instability and disruption (e.g., facing an enemy in war – Coles, Reference Coles2002), if at all, is such a simple and monolithic agreement about group identity likely to be found.

Moffett also defines societies as necessarily controlling some sort of territory, but while he is right to describe many social groups as “imagined communities” (Anderson, Reference Anderson1982), he pays less attention to the fact that many (and more every day) are also deterritorialized, where the ties between identity and space are weak (e.g., Appadurai, Reference Appadurai1990). So what to do, for example, about members of diasporic and other immigrant groups who hold dual identities? Moffett excludes them from membership in a society unless “there is little ambiguity about the point when each immigrant is accepted as a member.” Here, too, his view of how identity works could use further exploration, as his definition of society leaves out too much of the complex nature of human group identity formation, where maintaining both local and global, multiple, situational, and hybrid identities (as well as identity confusion) are all possibilities (e.g., Eriksen & Schober, Reference Eriksen and Schober2016; Friedman, Reference Friedman1994). Any definition of society must take this variability and flexibility into account.

Further, there is an inherent contradiction in defining society on the one hand as based on perception (shared identity) and on the other as objectively established (physical territory, “land or stretch of sea” that is under “absolute control”). If identity can be symbolically marked, then why cannot resources held by a society be symbolic as well, as in the case of ancestral homelands or pilgrimage sites not under groups' control, or otherwise disputed territory? Cultural knowledge too is a resource that, when shared, reinforces group identity, as for example knowledge of the properties of medicinal plants (Menendez-Baceta et al., Reference Menendez-Baceta, Aceituno-Mata, Reyes-García, Tardío, Salpeteur and Pardo-de-Santayana2015). Thus, perhaps “resources perceived to be held in common,” of which physical territory is one example, would be a better definitional criterion.

Finally, it is theoretically problematic to discuss group membership identification as either arrived at through genetic and psychological adaptations related to appearance, scent, and so on (“individual recognition societies”) or by means of culture and symbols (“anonymous societies”), with humans relying on the latter. In fact, there is a complex relationship between the two, and both can operate simultaneously in the same social group. For example, individual kinship recognition cues are routinely manipulated to reinforce ties among anonymous non-kin in institutional and other contexts (Qirko, Reference Qirko2013).

In short, identity is complicated: Individuals may adopt multiple, even conflicting identities, and “a shared identity does not have the same meaning for everyone who embraces it” (Sökefeld, Reference Sökefeld1999, p. 423). Moffett's wide-ranging paper is valuable in many respects, and particularly as regards the current “semantic mess” in operationalizations of social groups across species, but I believe his definition of a society requires more work before it will be fully useful.