David Pyrooz and James Densley’s recounting of reactions to their bold foray into the troubled waters of applied social science is instructive on several counts, in effect highlighting both the hazards and importance of the issues addressed to a variety of publics (political and social in the broadest sense, including the academic). Because space limitations preclude detailed examination of these issues, we focus primarily on their significance for the academy and for the publics to which we are ultimately accountable.

Conceptual issues arise quickly in Pyrooz and Densley’s article, beginning with identification of “Antifa” as a group at the forefront of a “budding social movement.” The authors then note that “factions within antifa were indeed durable across time, street-oriented, and youthful groups, and, importantly, intentional in their illegal behavior,” which is “central to their collective identity” – the Eurogang definition of gangs. Thus, despite the overall frame of their essay, their conclusion is that only some segments within Antifa are youth street gangs.

Doing away quickly with the first of two “sticking points” in defining gangs, i.e., that gangs have an organized structure, Pyrooz and Densley turn to the question of whether or not to include criminal activity in the definition. For several reasons, they affirm the Eurogang consensus to include it. First, they note, criminal and violent behaviors are not the only gang outcomes of interest, making it possible to define gangs in these terms while avoiding the tautology problem.

Second, Pyrooz and Densley argue that “the unit of analysis in this area of study is ‘gang members’ rather than ‘gangs,’ which has allowed researchers to skirt around problems of definition and tautology by instead studying individuals who are at liberty to self-nominate (or not) as gang members.” Although Pyrooz and Densley correctly observe that this focus on individuals has contributed to much being learned about the variability of behavior within gangs, we believe it is in important respects a fundamental limitation of gang research insofar as it fails to capture the dynamics of behavior as a product of intra- and inter-gang relationships, and of relationships with community institutions and forces occurring beyond local communities. As noted by Sierra-Arèvalo and Papachristos (2015: 157–158, emphasis in original), “It is exactly the enhanced groupness of gangs that differentiates them from common, passing delinquent groups, and which makes gangs analytically interesting. While gangs are made up of individuals, the life and culture of the gang is largely defined by group processes and resulting structures. Importantly, not only do relationships exist within gangs, but the relationships between gangs can also shape the structural reality of gang life.”

For many of us, a major point of contention is the last of Pyrooz and Densley’s justifications for including criminal behavior in the definition of gangs. Without it, they charge, street gangs are “conflated with many affinity groups.” Far from conflating relations among affinity groups, we regard study among them to be a rich subject of interest, theoretically and empirically. Antifa appears to be an excellent example, as the remainder of Pyrooz and Densley’s paper demonstrates.

Following discussion of these issues, Pyrooz and Densley jump to legal concerns, noting that statutory definitions of gangs converge on criminal behavior, that the “Guidance on STEP runs some 92 pages, most of which is legal precedent,” and that the law has withstood a variety of constitutional challenges.Footnote 1 While we recognize that control of crime and criminality are of primary social (including academic) concern, our interests have focused less on control, especially official control, than on advancing knowledge of conditions and processes associated with these phenomena. These include legal aspects of control, statutory certainly, but especially in practice, what happens on the ground, as experienced both by those who are most directly responsible for control and those who are subject to it.

After a thorough defense of the Eurogang definition of a gang, Pyrooz and Densley elaborate their Wall Street Journal op-ed to establish that certain segments within Antifa are violent, thus satisfying a key criterion of the Eurogang definition, political ideology notwithstanding.Footnote 2 They also argue that these segments constitute groups having collective identities “no less visible than a gang’s” and using signs and symbols to demarcate group boundaries. Moreover, consistent with the Eurogang definition, they find these groups to be street-oriented and durable, as indicated by “Antifa’s ongoing presence at rallies” throughout the country and public forms and targets of political protest. Finally, although they neglect to mention whether these groups are comprised of youth, they note in the original op-ed that gangs “tend to be composed of people in their teens and 20s, not unlike the age demographics of Antifa members arrested lately” (Pyrooz and Densley 2017).

In concluding that “Antifa checks every box on the gang list,” Pyrooz and Densley recommend proactive policing strategies commonly used with gangs – e.g., intelligence gathering, focused deterrence, civil injunctions – as more effective alternatives than the typical “crowd control” approach to far-right groups. We concur but hasten to add that, in addition to its relevance to law enforcement and crime control, study of Antifa is important because it provides an opportunity to extend knowledge of group and other collective processes beyond street gangs toward the discovery and demonstration of general human behaviors. Some recent work has been devoted to relating what has been learned about gangs to studies of other phenomena, such as internal conflicts within and between nations (see, e.g., Short and Hughes 2015; Hagan and Rymond-Richmond 2009; Savelsberg 2010). Antifa seems a likely candidate for such study. However, rather than focus on official (state and other jurisdiction) similarities between Antifa and street gangs, we would ask: In what ways is Antifa similar to and different from social movements, and how does that vast literature inform the nature of Antifa? By what processes did “factions within Antifa” become gangs – if that is what they are – and what is the nature of relationships and interactions between and among violent and non-violent segments? How did such differences develop and to what extent have they persisted over time? To what extent have segments within Antifa institutionalized?

Finally, a comment on research strategies seems appropriate. Flexibility in the design and execution of research recently has come into increased focus, and, in this respect, Pyrooz and Densley are to be commended. Quite aside from the technical and legal aspects of their argument, their kind of careful examination of phenomena to be included in research too often is foreclosed by premature and narrowly circumscribed criteria. Noting the advantages of a research strategy of retroduction, as distinct from both induction and deduction, Short and Strodtbeck (1965) are among a handful of scholars arguing that researchers should leave themselves open to alternative ways of collecting and interpreting data (Abbott 2016; McTaggart 1908; Peirce 1934). The basic premise is recognition of the everchanging nature of social life and the impossibility of fully capturing its reality by existing methods that are necessarily arbitrary and fixed in time.Footnote 3 This is no less true of Antifa than of youth gangs, regardless of whether or not they come to be defined as one and the same.