Like Baumeister et al., for more than 10 years we too have asked the question: Are groups more or less than the sum of their individual parts? Whereas Baumeister et al.'s answers mostly are informed by social psychology, ours are informed by experimental cognitive psychology (the “collaborative recall” literature; e.g., Barnier et al. Reference Barnier, Sutton, Harris and Wilson2008; Harris et al. Reference Harris, Paterson and Kemp2008), by philosophical ideas about “distributed cognition” (where philosophers argue that human cognition extends beyond the mind of an individual to incorporate external resources; e.g., Sutton et al. Reference Sutton, Harris, Keil and Barnier2010), and by memory studies in the humanities. Our research aims to draw links between the remembering practices of one person and the remembering practices of two people, three people, groups of people, and communities, cultures and nations. From this perspective, we comment on three aspects of Baumeister et al.'s ambitious and provocative article.

First, we are less convinced than Baumeister et al. that differentiation of selves predicts “many of the best outcomes of groups” (abstract) and that “most of the bad effects of groups … come when the individual self is lost or forgotten as identity is submerged in the group” (sect. 4, para. 1). We agree that “relationships among individuals are not fully reducible to properties of the separate individuals” (para. 2). However, we do not agree that differentiated roles within social groups necessarily “make more powerful and effective systems” (para. 2). In our interdisciplinary research program on collaborative remembering by strangers, friends, family groups, organizational teams, and long-married couples, we find that successful group performance depends on a balance between both integration and differentiation of knowledge, information, and/or expertise of individuals in groups (Wegner Reference Wegner, Mullen and Goethals1987), as well as a balance between individual and group identity and individual and group cognitive strategies and processes. Indeed, individual (vs. group) focus and strategies often lead to the poorest group outcomes (e.g., Barnier et al. Reference Barnier, Priddis, Broekhuijse, Harris, Keil, Cox, Congleton and Addis2014; Harris et al. Reference Harris, Keil, Sutton, Barnier and McIlwain2011).

In section 2.1.3., Baumeister et al. briefly introduce Wegner's (Reference Wegner, Mullen and Goethals1987) “transactive memory theory” (para. 3). Wegner proposed that people in well-established groups share encoding, storage, and retrieval of information such that groups may perform better than the sum of their individual parts and may show “emergent” outcomes. Baumeister et al. describe a meta-analysis and three studies highlighting differentiation, drawn from the large transactive memory literature in organizational psychology. This literature has blossomed recently with contributions on collaborative recall (from cognitive psychology) and distributed cognition (from philosophy) (e.g., Barnier et al. Reference Barnier, Sutton, Harris and Wilson2008; Hirst & Echterhoff Reference Hirst and Echterhoff2012). We believe there is even more insight to be gained from transactive memory theory, especially from research with the kinds of intimate family groups that Wegner developed his theory to explain, where successful group processes (Baumeister et al.'s second step) may depend on a balance between differentiation and integration. That is, where “I” and “we” operate in harmony to predict both group and later individual performance (e.g., Harris et al. Reference Harris, Keil, Sutton, Barnier and McIlwain2011).

Second, any account of the success or failure of groups needs to consider and explain the huge variation in constellations of people and processes that take place within and across groups. Baumeister et al. argue that a two-step process predicts positive group outcomes. The implication is that in research studies, as in life, all groups that follow these two steps will be successful, whether in a group of three people recalling a set of information, a team building a device, or a committee making a decision. But underneath group-level effects (where groups that collaborate on average perform differently from groups that do not collaborate), we have noticed substantial differences in the ways in which groups of the same type operate, which seem to reflect meaningful yet unexplained variation in group membership, function, dynamics, strategies and/or something else (e.g., Harris et al. Reference Harris, Barnier, Sutton and Keil2014a).

Third, understanding group performance requires, we believe, nuanced measurement (of the sometimes idiosyncratic profiles) of what individuals bring to, do in, and get out of groups; where the unit of analysis sometimes is the individual and sometimes is the group. Consider, for example, how Baumeister et al. (and the broader group-processes literature) count superior group performance: as better or worse than “the sum or average of their parts” (para. 5). Sums and averages, however, can miss the subtleties of what's going on in groups. In our research, we have asked long-married, elderly couples to recall names of their families and friends either alone or together (e.g., Harris et al. Reference Harris, Keil, Sutton, Barnier and McIlwain2011). Imagine couple 1, who together recall lots of names, one name after another, rapidly cross-cuing each other with names. The husband and wife each recall 10 names in this way. Now imagine couple 2. The husband recently suffered a stroke and has memory problems. He recalls 4 names and his wife recalls 16 names, scaffolding their joint recall and dominating the conversation. For both groups, their sum is 20 and the average is 10, but discrepancies across individuals (e.g., in ability or expertise) within these same kinds of groups tell a richer story of the individual and joint processes that give rise to the group's performance.

Finally, Baumeister et al. might consider more the broader functions or goals of group performance. It is crucial to understand what individuals and groups themselves explicitly or tacitly count as success or failure. In the domain of memory, for example, simply remembering more or remembering correctly (as measured by standard amount and accuracy outcome variables) often is not as important to people as the broader functions of telling and sharing stories, including building individual and group identities and understandings (Harris et al. Reference Harris, Rasmussen and Berntsen2014b).

Exploring when and how “I” becomes “we” is crucial because in everyday life, across our lives, we engage in many cognitive activities – such as remembering – in the presence of, prompted by, and in partnership with others (Barnier Reference Barnier2010). Complementary, albeit challenging, views from across psychology and across disciplines can help us understand the relationship between the cognitive and social lives of individuals and groups.