Abstract
This critical review of more than three decades of studies of frames and framing in Social Movement (SM) research first offers a brief history of the notion of ‘frame’ in various disciplines, and then discusses empirical studies of frame alignment, frame disputes, frame resonance and master frames, among other notions. It is found that the very notion of discursive of cognitive frames remains very vague in these studies, and what are actually studied are for instance, beliefs, attitudes, goals, ideologies or values, especially how they are expressed in discourse. Also studies of relations between frames and culture, identity and discourse show that the notions of frames and framing are theoretically and methodologically unsatisfactory in empirical studies. It is concluded that the methods of the cultural paradigm of SM research may take advantage of the advances of more explicit methods in the study of language, discourse, interaction and cognition.
Introduction
More than 30 years have passed since the first framing studies in Social Movement (SM) research were published. The notions of frame and framing have become very popular in these three decades, especially in the ‘cultural’ paradigm of social movement studies. This paper offers a critical review of some of these studies. It does so especially from the perspective of contemporary studies of discourse, because it precisely among the most important notions that define the cultural approach in social movement research.
The review will show that generally the notions of ‘frame’ and ‘framing’, as used in this research paradigm, are very vague and ill-defined theoretically and therefore also methodologically inadequate. In most studies they serve as a fuzzy term to refer to a large variety of discursive and cognitive phenomena that have been studied, in several disciplines, with much more precise concepts and methods.
It is concluded that given the limited theoretical and methodological value of the notions of frame and framing, cultural social movement research should abandon using these notions, and make use of more precise theoretical and methodological concepts of the phenomena it is studying, for example, as they are offered in neighboring disciplines.
The critical review of the notion of frames and framing in social movement studies, does not imply that these studies are less interesting or irrelevant. On the contrary, as sociological studies of social movements they are very important, but not because of the use of the notions of frames and framing, but because of
Corpus of articles reviewed
For this review, 162 articles have been reviewed. The main selection criterion was that articles should have the words frame or framing in their titles and study social movements. Most of these studies are in sociology, and some in social psychology and political science. The words frame(s) or framing occur 20,350 times in these articles (125 times per article).
The relevance of discourse in social movement studies of framing is obvious from the fact that 116 articles feature the word ‘discourse’ – with a total frequency of 2153 occurrences (in 46 articles the word ‘discourse’ appears more than 10 times). Because of space limitations, some of these discourse analytical studies in social movement research will be reviewed in a book on social movement discourse in preparation. The same is true for the use of ideology and other sociocognitive notions used in social movement research (the word ‘ideology’ is used 1047 times in 121 articles on social movements). These discourse and ideology articles will only be briefly mentioned in this review.
Theoretical framework: Multidisciplinary Discourse Studies
Since most framing studies use some kind of discourse as their empirical data, this critical review is written against the background of contemporary discourse studies, as it has been developing since the 1960s (for detail, see, e.g. Angermuller et al., 2014; De Fina and Georgakopoulou, 2020; Gee and Handford, 2012; Tannen et al., 2015; Van Dijk, 2011). Relevant for the study of social movements is especially the research in Critical Discourse Studies (CDS)(for a recent selection, see Flowerdew and Richardson, 2018).
More specifically, the criticism presupposes the theoretical and analytical framework of my sociocognitive approach to discourse and its relations to (anti)racism, power, ideology, context and knowledge. This framework is particularly relevant for the critical review of framing research, because it precisely focuses on the two crucial notions used in framing research in the cultural paradigm: discourse and cognition (see, e.g. Van Dijk, 1977, 1993, 1998, 2008a, 2008b, 2009, 2014, 2021: for an introduction, see Van Dijk, 2015). A more focused use and explanation of this theoretical framework is relevant in the critical review of extant discourse analytical approaches in social movement research briefly referred to below. In his article, this framework only serves as a general background for the criticism of framing studies. Space limitations does not allow detailed explanation of some of the theoretical notions mentioned below (see the books just cited, and a future book on social movement discourse).
A brief history of the notion of ‘frame’
As we shall see in more detail below, SM research generally attributes its concept of ‘frame’ to Irving Goffman’s book Frame Analysis (Goffman, 1974). The notion, however, has been used before in several disciplines. This history and uses, partly referred to by Goffman himself, may be summarized as follows.
Goffman refers to
Goffman also refers to the work on the ‘common-sense world’ and ‘multiple realities’ by
One of the definitions of framing is in terms of defining the situation. Not surprisingly, Goffman also refers to the classical definition of Thomas (1928) about the reality of social situations (and their consequences) if people define them as real – though he adds that these consequences may well be very mundane, and hardly noticed by social actors. This perspective of framing also defines the very philosophy of social constructionism in several social sciences (Hjelm, 2014).
The link with contemporary micro-sociology, ethnomethodology and conversation analysis is established by Goffman’s reference to the work on multiple realities by Garfinkel (1967), who stresses that the meaningfulness of everyday activity ‘in a seemingly unlimited number of situations’ (p. 5) is reducible to a small set of rules and practices. Deviation from such rules would thus explain how the intelligibility of deviant acts is undermined – a concern that inspires much of the theory and the analyses of the many forms differently ‘keyed’ actions and situations in the rest of Goffman’s book.
More specifically Goffman (and others) attribute the notion of frame to Bateson (1955/1972), who uses the notion as a definition of the communicative or interactive situation or ‘key’, as when animals know whether some action should be interpreted as serious or as ‘play’. Thus, frames for Bateson are interpretation schemas, which has become the standard (cognitivist) definition of the term in much SM framing research.
Goffman’s book was published in a year (1974) in which also other influential studies were published such as the first study on conversation by Sacks et al. (1974) on turn-taking, initiating the vast field of Conversation Analysis (Sidnell and Stivers, 2012), for which also Goffman himself was a major inspiration (indeed, Schegloff did his PhD with Goffman).
Two years later, Fillmore (1976), published his seminal paper ‘Frame Semantics and the Nature of Language’, which would give rise to a broad research paradigm (Frame Semantics) on frames as a basis especially for the study of word meanings, and later more generally research in the field of cognitive linguistics (Croft and Cruse, 2004; Evans and Green, 2006; Fried and Nikiforidou, 2013; Littlemore and Taylor, 2014). Like Goffman, Fillmore was interested in framing as ‘the appeal, in perceiving, thinking, and communicating, to structured ways of interpreting experiences’. Later studies of the semantic frames of concepts (as expressed in words) further elaborate the knowledge structures underlying these concepts.
Fillmore refers to the notions of ‘frame’, ‘schema’ and ‘scenario’ as they have been introduced, more or less at the same time, in Artificial Intelligence, for instance in an influential article by Minsky (1975), published the year before, in order to account for structured knowledge of the world. This work was itself inspired by Quillian’s work, in the 1960s, on the representation of conceptual information (Quillian, 1963). Minsky’s classical example is that our knowledge of tables, in the form of a frame, allows us to know that tables have legs even when we don’t see them.
The notion of ‘schema’ more generally characterizes the cognitive revolution of the 1960s and 1970s (see, e.g. Neisser, 1967), inspired especially by Bartlett’s (1932)
A year after Fillmore’s paper,
My own book Text and Context (Van Dijk, 1977) introduced the notion of knowledge frames in systematic analyses of discourse, for example, to define local and global coherence, the relation between discourse and the world, as well as the notion of context. Beyond the obvious influence of psychology and AI, this book also was inspired by model theories in logic, for example, as a basis of formal semantics, and thus prepared the way for the introduction of mental model theory as a basis for discourse processing (Van Dijk and Kintsch, 1983; see also Johnson-Laird, 1983). We shall come back to the crucial notion of mental models below. Later these notions also influenced the studies of the perception and comprehension of events (e.g. Zacks and Tversky, 2001).
In a collection of studies by her students and colleagues, discourse analyst Tannen (1993) explicitly refers to ‘frame analysis’ in terms of structures of ‘expectations’ in text and talk, based on different kinds of knowledge structures. After a detailed summary of earlier studies of frames and related notions (schema, script, etc.) she reports on her own study of framing, as part of the well-known ‘pear stories’ film shown to subjects from various countries (see Chafe, 1980), of which Tannen analyzes the Greek data.
We see that especially in the 1970s there were developments in several disciplines of the humanities and social sciences to account for the ways situations, events and ‘worlds’ were discursively and cognitively represented in terms of schemas, frames, models or scripts. Especially in various areas of cognitive science, these notions also have been formulated in (more) explicit terms.
Goffman’s frames
Since SM research to frames is indebted especially to Goffman, and – regrettably – less to the other studies and disciplines dealing with frames mentioned above, let us also summarize some of Goffman’s ideas. He formulates the aim of Frame Analysis (1974) as follows:
(. . .) to try to isolate some of the basic frameworks of understanding in our society for making sense out of events and to analyze the special vulnerabilities to which these frames of reference are subject (p. 10).
Rather surprisingly, given his fundamental interest in framing, both in this book, as well as in his other books, Goffman is quite succinct when it comes to the theory of the structures or rules defining such frames. He distinguishes, first of all between natural and social frames. The first are used to understand natural events, such as the weather, and the second to understand events and actions (‘guided doings’) in which human beings are somehow involved (as would be the weather forecast), and subjected to standards of social appraisal (e.g. of honesty or efficiency), on the one hand, or the constraints of nature, on the other hand – thus also giving rise to two kinds of situations: manipulations of the natural world, or interactions in the social world. These ‘primary frameworks’ allow the users to ‘locate, perceive, identify, and label a seemingly infinite number of concrete occurrences’ defined in its terms. Primary frameworks vary by their degree of organization. Some are neatly organized by entities, postulates and rules, but most others appear not to have an articulated shape (p. 21). In the same situation, several frameworks may be applied, as when a waving hand can be interpreted as a greeting or as a summons to approach. Frames for Goffman, however, are not only the rules or premises that govern an activity, that is, as a form of cognitive control, but also may be described as the very organization of the activity itself, because the actors fit their actions to the understanding of the activity (p. 247). We shall see later that this ambiguity of the conception of frames and framings, viz. between mind and action, is quite characteristic also of the use of frame theory and analysis in the study of social movements.
At this point, Goffman hardly becomes more specific about the organization of these primary frameworks. Crucial is to distinguish, at this point, the definition of a frame as an individual, subjective interpretation of a social situation (defined below as a mental situation ‘model’), on the one hand, or the socially and culturally shared system of knowledge or rules applied in the understanding of social situations, on the other hand. Goffman explicitly opts for the second definition when he associates primary frameworks with social groups and their culture, for example, in terms of ‘principal classes of schemata, the relations of these classes’ (p. 27). In this case he refers to belief systems or ‘cosmology’ – notions about ‘cognitive resources’, he says, which scholars of social life prefer to leave to others. Indeed, we’ll find below that this is indeed the case in much of the studies of frames in the sociology of social movements.
In the rest of this book, instead of elaborating a more explicit theory of frames, Goffman gives a vast number of examples of how in the real world and in language events they may be interpreted in different ways depending on their ‘keying’ or ‘framing’, for example, as play or serious. He is especially interested in the many (re)transformations involved, as in fabrications, errors, malfunctions, illusions, and ambiguities in everyday situations. Especially in his frame analysis of talk, he stresses that all talk (and storytelling, etc) is a form of performance, as on the stage. He concludes:
Realms of beings other than the ordinary provide natural experiments in which a property of ordinary activity is displayed or contrasted in a clarified and clarifying way. The design in accordance with which everyday experience is put together can be seen as a special variation on general themes, as ways of doing things that can be done in other ways. Seeing these differences (and similarities) means seeing (p. 564).
We see that although SM research attributes the notion of frame mainly (though not only) to Goffman this indebtedness is quite superficial, and in fact limited to a notion of frame as an interpretation schema of social reality. The huge variety of special (illusion, play, etc.) frames studied by Goffman have hardly any influence on later framing studies of social movements.
Frame analysis in social movement research
A decade after the publication of Goffman’s seminal book, sociologists in the field of social movement research continued the same phenomenological tradition of symbolic interactionism and related approaches by studying collective action in terms of frames, and as interpretations and definitions of social situations. Despite considerable theoretical and methodological differences, as well as disputes about the nature or relevance of these notions, frames and framings were generally adopted as fundamental notions accounting for the role of interpretations of the experiences of the participants of social movements. Some of the notions used in SM research based on a concept of ‘frame’ will be briefly summarized here.
Frame alignment
In one of the first and most cited studies on frames in SM research, Snow et al. (1986) introduced the notion of
The authors distinguish between
The notion of frame in the study of social movements in the paper by Snow, et al. is motivated by the criticism of earlier approaches using the notion of grievances, with the argument that grievances as such are not the relevant cause of collective action, but how such grievances are interpreted by social actors. This important constructionist thesis defines the cultural paradigm of social movement studies and is also shared by my sociocognitive approach in discourse studies, which claims that the relations between discourse structures and societal structures are cognitively mediated.
The theoretical task to be conceptually made explicit is to specify as what exactly such grievances are interpreted, and especially how such an interpretive process takes place. This would require detailed cognitive analysis of frames, an analysis regrettably not provided by the authors, nor by the authors of most studies of frames, despite the fact that contemporary cognitive science offered many concepts to do so.
Such a more detailed cognitive analysis would also be required to describe the congruence relating personal frames with SM frames: one would need a cognitive definition of this relationship, but such a definition presupposes an explicit cognitive description of frames in the first place. The same is true for the definition of the various kinds of frame alignment, such as extension, amplification and transformation.
Moreover, what is also ignored here and elsewhere is the important role of context in such processes, because an experience no doubt is interpreted (‘framed’) in a different way by different social actors in different social situations.
As we shall see also below for most other studies, the main problem with the introduction and application of the notion of frame, borrowed from Goffman, is not only that it is not made explicit in terms of mental representations but that it glosses over many very different notions that hardly can all be summarized as forms of ‘interpretation frameworks’, such as beliefs, ideologies, goals, interests, values or activities. Each of these notions need specific theoretical analysis in their own right, and vaguely ‘summarizing’ them as ‘frames’ does not contribute to our insight in the interpretative aspects of social movements. In sum, one would expect a detailed account of the cognitive (and discursive) processes of how people go about understanding, comparing, negotiating, accepting, rejecting, etc. the beliefs, goals, ideologies, etc. of social movements in relation to their own. Vaguely defining or ‘summarizing’ these as ‘frames’ does not contribute to the precision of the analysis.
The authors criticize that earlier network approaches ignore
what transpires when constituents and bystanders or adherents get together. Since a good portion of the time devoted to many SMO activities is spent in small encounters, an examination of the nature of those encounters and the interactional processes involved would tell us much about how SMOs and their constituents go about the business of persuading others, effecting switches in frame, and so on (p. 468).
Now, that would indeed be a relevant sociological task, but that is regrettably what the authors do not provide. They give some brief examples of participant talk, but they do not actually analyze such talk, despite the existence of more than a decade of Conversation Analysis in sociology itself (Sidnell and Stivers, 2012), of interpersonal persuasion research in social psychology (Dillard and Shen, 2013), or of the study of argumentation (Van Eemeren et al., 2014) in discourse studies.
What is problematic at the theoretical level unfortunately cannot be resolved at the empirical, analytical level. For instance, an example of frame bridging is described in terms of the use of direct mailing of different conservative Social Movement Organizations (SMOs) in the USA – which is rather a form of discursive organizational activity, and hardly a form of frame or frame bridging in the first place.
In another illustration, based on Benford’s research (see below), the Texas peace movement tried to mobilize supporters by emphasizing democratic values of free speech, besides very general values such as justice and cooperation. In other words, what happens here is the use of persuasive, mobilizing discourse focusing on or emphasizing general values, shared by the whole community, instead of the more specific and possibly less supported ideological values or goals of the peace movement. Again, in order to describe and analyze what happens in these processes of mobilization is better described in terms of the structures of persuasive discourse and of ideological values, and not in the much vaguer terms of frames of frame bridging.
The same is true for the example (based on Snow’s own research) of the appeals of Salvation Army persuading local residents opposing a center for homeless men, a resistance that can be accounted for in the more specific terms of prejudices and ideologies. Also here, the notion of frame is much too general and vague to describe the cognitive and social processes involved. In fact, in these and most of the other examples of this article the notion of frame soon disappears in the analysis – showing that it is analytically superfluous.
In a review of 25 years of frame and frame-alignment research since the publication of this article, Snow et al. (2014), unfortunately fail to repair the theoretical vagueness of the earlier paper of 1986. One of the authors, Holly McCammon, admits that some aspects of the original theory, such as the specification of the actual processes or the relations with other notions, such as ideologies and schemas should be developed in future research.
Other studies of frame-alignment
After this early theoretical article of Snow et al. (1986) the notion of frame-alignment has been used in many SM studies, of which some will briefly be mentioned here, without further analysis, because they show the same theoretical and methodological limitations. This does not mean that these studies are not interesting as sociological studies of social movements, but generally not because of the use of the notions of frame or framing, but rather because of the analysis of activities, organization and mobilization. My main focus, in the review of these and other SM articles on framing, is to show what other notions are actually used, instead of frames, in order to describe a social movement and its activities, and why therefore the notion of frame is superfluous in these studies and/or glosses over more specific phenomena.
Instead of discussing the other papers on frame alignment individually, and repeat the same critique again and again, we summarize these studies in a simple schema (see Table 1), in which we mention the overall topic, and what aspects of social movements is described without further specific analysis, and hence glossed over, as frames.
Frame alignment studies and the notions glossed over as frames.
Frame disputes
In an extensive and influential case study of the Texas peace movement, briefly mentioned above, Benford (1993) examines internal disputes between radical, liberal and moderate factions of the movement, based on 2100 hours of fieldwork, nearly 1000 pages of field notes, dozens of interviews with core activists and 132 informal interviews with rank and file members. We see that his study is based on a vast corpus of text and talk, but such discourse is not analyzed in terms of contemporary discourse analysis, but in terms of traditional content analytical ‘coding’ of the various discourse genres collectively defined as a ‘dispute’. Such a dispute not only is expressed in various structures of text, talk and interaction, but also represented in different kinds of social cognition, such as knowledge, attitudes and ideologies (Benford here uses the notion of ‘ideological contours’), two complex levels of analysis here subsumed under the notion of frame.
The frame disputes studied pertain to what generally in SM framing studies have been called diagnostic, prognostic and resonance frames. Diagnostic frames are defined in terms of the identification of a problem or the attribution of blame or causality (p. 686). A further cognitive analysis of the structures or kinds of mental representations of these frames, however, is not provided. Thus, a dispute may arise about the emphasis to give to which of various causes of a problem, for example, the nuclear threat itself, its possible causes, or more generally on other important social problems, such as poverty.
We see that the dispute is not so much about the frame (the definition of the situation), but about the discourse about the social situation: what to emphasize or not – which are semantic or a rhetorical aspects of discourse (see below for frames in media research). Unfortunately, and typically, the actual frames are not but vaguely characterized in the article, only the frequencies of types of disputes between the three factions, as well as some discourse examples expressing the dispute. Consider, Benford’s informal characterization of the radical ‘perspective’:
From the radical perspective, the nuclear threat was but a symptom of global systemic problems exacerbated to a large extent by the U.S. and its pursuit of hegemony as well as a manifestation of greedy defense contractors, multinational corporations, and unscrupulous politicians. Radical SMOs called for fundamental structural changes, such as a basic redistribution of resources globally, changes they contended would make war obsolete. Hence, this radical wing identified and articulated links between various social problems - poverty, oppression, militarization, and environmental misuse. Finally, these SMOs supported the use of, and occasionally employed, unruly tactics, including civil disobedience and other forms of nonviolent direct action. (p.681)
Notice that this characterization of the frame (here described with the equally vague and undefined notion of ‘perspective’) is informally derived from the interpretation of many different discourse genres of the (members of the) radical faction, produced in different communicative situations, also involving different audiences, different goals, etc. The actual dispute does not take place in terms of this reconstructed frame but in terms of the actual discourses of the faction. Secondly, as a sociocognitive representation, the ‘Nuclear Threat’ is a socially shared attitude of the radical faction of the peace movement, in this case based on a more general anti-imperialist ideology, and related to other attitudes (e.g. about poverty, oppression, etc.).
Each of these could then be structurally analyzed, for example, in terms of causes of the problem attributed to participants with different identities and roles (such as the U.S., politicians, defense contractors, corporations), and their properties (unscrupulous) their actions and their goals (hegemony), and the ways these should be opposed (civil disobedience, etc.). Since most of the disputes are found to involve the ‘ideological wings’ of the movement (p. 685), we actually need a more specific ideology analysis, and not a frame analysis, to describe and explain the data.
Similarly, to describe the extension of the concerns of the movement beyond nuclear disarmament, for instance to feminist issues and anti-gay rights, the author more specifically uses more specific notions such as goals and values (p. 688) and no longer uses the notion of frame.
Although most of the disputes in the Texas peace movement are about the various aspects of the ‘definition of the situation’, there were also disputes about how ‘reality should be presented so as to maximize mobilization’. In this case, we don’t only need a specification of underlying forms of social cognition, but a detailed analysis of discourse structures, such as the expression of overall topics, actor descriptions, argumentation and other persuasion strategies, including the use of specific words instead of others, metaphors, stories and so on. The credibility of the speakers plays an important role here, and such a parameter requires an analysis of the pragmatic structure of the communicative situation, specifying social properties (identities, roles, relations, status, power, etc.) of the participants, as represented in the context models of the participants (see Van Dijk, 2008a, 2009).
Ruffa and Sundberg (2018), in a quantitative (questionnaire) study influenced by Benford’s study of frame disputes, examine the negative effects (e.g. of cohesion and performance) of such disputes of individual soldiers ‘breaking’ the dominant frame of the military organization, defined as the humanitarian ‘mission’ during peacekeeping actions in Afghanistan. While interesting as a sociology study of personal and organizational aspects of the military, the notion of frame here involves many (concepts of various) aspects of military action in terms of the soldiers’ definition of the Italian contribution to the NATO mission in Afghanistan, for example, as ‘war’, ‘peacekeeping’ or as ‘humanitarian intervention’. Since many aspects of military action, tasks, war or peacekeeping are involved, the notion of frame here seems to summarize the various aspects of the perception and evaluation of the mission, including individual experiences and even emotions. In other words, the notion of ‘frame’ in this study does not contribute anything analytically relevant beyond the experiences and opinions of the soldiers, as compared to the dominant military or political ‘views’ of the mission. A discourse analytical study of interviews with soldiers may have contributed more insights in the detailed ‘frame’ structure of their ‘definition’ of the mission.
Frame variation
Snow et al. (2007), in a vast comparative study of the media coverage in various countries of the 2005 French Riots, deal with frame variation. Instead of detailed discourse analysis of this coverage, also here traditional content analysis is used to deal with so many data. Framing is informally defined as in most earlier studies of Snow and others, namely as ‘interpretive’ process (which would be a mental process). As elsewhere, the hardly helpful metaphor of the picture frame is used (as well as the metaphor of being in and out of focus), how events are defined, for example, in terms of their causes or how they can be controlled. However, in this case, the framing is obviously discursive, because the data are newspaper articles, and the aim is to show, by comparative international research, how news (discourse) frames may vary – a type of variation little studied before (see below for ‘frames’ in research on news media).
The authors do so by examining four types of frame variation in the coverage: ideology (here of country government and newspaper), sociopolitical context (e.g. unemployment in the country, or distance from the country), attribution (who is blamed), time (frames change after some time). These conditions of variation, unfortunately, are not theoretically analyzed, but informally postulated (or derived from the informal analysis of the news items).
One of the problems of this study, as well as of many other framing studies in SM research, is the lack of a more sophisticated multidisciplinary theory for a social constructionist approach. Thus, the crucial point here, overlooked by the authors, are the relevant dimensions of news production (see, e.g. Van Dijk, 1988). The variable conditions mentioned are only relevant when they are known and defined, that is, cognitively mediated, as such by the actual journalists. The ideology of a government cannot possibly be a direct condition of discursively expressed frame variation, unless such an ideology is adopted by the newspaper and journalist. Similarly, sociopolitical context of whatever type is relevant only when known to the journalists of these newspapers. Instead, frame variation is described in terms of how content of the news articles is coded, such as the diagnostic nature of the riots in terms of ‘riff-raff’, ethnicity of actors, over-reaction of the authorities, or their prognostic aspects, in terms of law and order, action program, better housing or limit immigration, and the source (e.g. politicians) of these framings, as well as their salience in the coverage. The results show that ideology and proximity hardly influence framing; time obviously is of influence (for instance on the crystallization of a frame), whereas the influence of the actors/speakers defining the events are crucial, especially in terms of their attribution.
Unfortunately, the method of this study, quantitative content analysis and coding of content, doesn’t show how exactly in the news articles the events are described, and how exactly news actors/speakers comment about them. That is, we never come to see how exactly journalistic mental models of the riots (informally characterized as ‘frames’) are expressed in the news text. We only get to know how many newspapers, in what country, and of what ideological orientation define or blame the riots in terms of riff-raff, ethnicity or government actions or ideology, among other causes – as coded by the researchers. Structurally this means a (content analytical, coded) definition of higher level semantic macrostructures of the articles, and not of the detailed, local way events are described or explained. Precisely that local variation of the news discourse in different newspapers in different countries would be interesting to study, as an expression of possible ‘frame variation’. Theoretically, such an analysis would also need to be described in terms of the ideologically and contextually controlled mental models of the events of the journalists as well as their sources. Content analytically based frame variation gives a very general picture of variation, but glosses over the more detailed discursive and cognitive processes involved in variation of newspaper coverage.
Other studies of frame variation, such as those of Krogman (1996) on environmental controversies, and of McCammon (2012) on demands of women in the jury movements in the early 20th century, also would be more explicit if instead of the notion of frame, more precise notions such as attitudes, rhetorical strategies, gender ideologies and argumentation structures were further specified.
More recently, framing studies are increasingly carried out in a discourse analytical perspective (19 papers mention ‘discourse analysis’ more than once, of which 9 are published after 2010). Thus, in a study briefly defined in terms of ‘frame variation’, Ravazzani and Maier (2017) examine opposing positions of websites about the Mauna Kea Mountain in Hawaii. The opposing ‘frames’ are defined in terms of different discursive strategies of legitimation as defined by critical discourse analyst Van Leeuwen (e.g. van Leeuwen, 2007), such as authorization, moral evaluation, and rationalization, both in discourse as well as images. They conclude that ‘the primary frames are: the sacredness and protection frame on Protect Mauna Kea website and the harmony and future frame on Maunakea & TMT website’. The question however is why the interesting use of multimodal discourse analysis in the comparison of two websites needs the notion of frame, when the authors themselves use the less technical notion of perspective. The interesting aspect of the article is not the notion of frame but how legitimation and other notions are applied in the analysis of competing websites.
Mobilization frames
Among many other functions, frames are developed to mobilize adherents, as was the aim of frame bridging in the Snow et al. (1986) article reviewed above. Thus, well-known Italian SM researcher Diani (1996) studied the relation between such frames and the political opportunity structure in Italy, especially with respect to the regional populism of the Northern League, a right-wing movement. The study of this movement is especially interesting because most SM studies deal with left-wing, progressive movements.
Diani combines a more classical approach of the study of political opportunities (political alignments, formal channels of access, allies and inter-elite conflict), on the one hand, with the more ‘constructionist’ approach in terms of different kinds of mobilization frame, which may vary as a function of political conditions, such as ‘realignment frames’, ‘inclusion frames’, ‘revitalization frames’ and ‘antisystem frames’, on the other hand (p. 1056).
Unfortunately, these frames are not spelled out in detail, but informally described in terms of the discourses that articulate them. These are fundamentally different things, also because discourses as forms of social interaction are always adapted to the communicative contexts: the ‘same’ frame may be expressed in very different discourses in different situations. Thus, realignment frames ‘emphasize the need to restructure political systems’. Inclusion frames are defined as rhetorical devices emphasizing the aspirations of new political actors, and antisystem frames are characterized by a mobilizing ‘message’ that challenges the fundamental traits of the political system. Revitalization frames are merely described as ‘reflecting the fact’ that systems must be revitalized from within.
Most of these definitions are brief informal characterizations of the (main) content of mobilizing discourse, and not as detailed cognitive or discursive representations. As is the case for many authors who use many different notions to informally ‘define’ frames, the author explicitly claims ‘I view frames here as abstract forms of political rhetoric rather than as belief systems anchored to specific contents’ (p. 1058). Yet of the discourses only the general rhetorical aspect of emphasis is mentioned, but such emphasis might be made by main topics, argumentation, the lexicon, metaphors or various forms of semantic and rhetorical enhancements. Although frames are thus vaguely associated with discourse structure (emphasis), no detailed discourse analysis of concrete examples is provided beyond informal comments on the main ‘themes’ of populist League discourse, such as their identity and their enemies.
The success of the League was thus explained in terms of the consistence of their mobilization message with a widespread anti-system master frame (see below). But in this case, the master frame obviously cannot be formulated in terms of discourse structures, but would need a more cognitive formulation in terms of shared attitudes and ideologies.
As is the case for most framing studies on social movements, also this interesting article offers a political explanation of the success of the Northern League, not only in terms of the political situation in Italy, but especially in terms of theoretical concepts (such as inclusion, anti-system) found in the themes (rhetorical emphasis) of the discourses of the League. But instead of offering a detailed discourse analysis (beyond referring to earlier content analysis), this explanation is formulated in terms of frames, whose contents or structures are not specified, nor the ways they are related to what is actually studied: discourse. Substitute the term of ‘frame’ by that of ‘discourse’ and the study would more concretely deal what it actually studies. No additional theoretical or empirical insight is provided by the use of the notion of frame.
Incidentally, few (14) other SM scholars use the notion of ‘mobilization frame’ – each just once (Benford, 1997, twice) and Diani 12 times, so the particular type of frame he uses is not very popular among SM scholars.
Master frames
A very popular notion introduced in the SM literature on frames is that of master frames, that is, a more generic frame that can be used by various SMOs in different ‘protest cycles’ and a broader context (Snow and Benford, 1992). It is used in half (76) of all articles. Stanbridge (2002) uses the notion 121 times in her article on master frames.
A distinction between microlevel vs. macrolevel (or master) frames may be theoretically sound, and reflects other ideas in sociology (between micro and macrosociology, see Alexander, 1987) or discourse studies (see the multidisciplinary book on macrostructures in discourse, action and cognition: Van Dijk, 1980), as well as many other theories in different disciplines. But, as is the case for frames, we would also need an explicit theory of master frames (e.g. in discursive or cognitive terms) and how micro and macroframes are related. Unfortunately, this is not the case in any study of master frames. In the review of some studies on master frames we’ll see that the notion of master frame is another vague concept that is used to describe many different discursive or cognitive structures, such as general values (justice), ideologies or goals, among others.
In their foundational article on master frames Snow and Benford (1992) first of all repeat their general theory of frame, for example, as active, process derived ‘signifying work’ of movement actors engaged in reality construction (p. 136), and hence as an ‘interpretive framework (. . .) that simplifies and condenses the world out there’ (p. 137). With this summary of the theory of frames and framing, master frames are defined as having the same functions as frames ‘but only on larger scale’ (p. 138). They do not characterize single SMOs, but SMOs more generically. They are like paradigms instead of specific theories. For the authors they are ‘analogous to linguistic codes in that they provide a grammar that punctuates and syntactically connects patterns or happenings in the world’. (p. 138) – a comparison that does not make sense at all, in any of the terms (codes, grammar, syntax) used. If making a ‘linguistic’ comparison at all, they should be defined as overall semantic macrostructures or topics in relation to the local meanings (microstructures) of words and sentences of discourse (Van Dijk, 1980).
Unlike more specific frames of specific movements, such master frames are so general and flexible that several movements can use the same frame. For instance, the successful Civil Rights frame of the 1960 in the USA was later adopted also by the women’s movement and the gay movement, among others. According to the authors other general master frames would be for instance, an injustice, oppositional or imperial master frame.
As defined in this study, in master frames also the attributional aspect is more general, for instance in terms of unjust external structural arrangements, and not a specific enemy or opponent of a specific SMO, as is the case for the use of the Civil Rights master frame by many different SMOs. Again, a comparison is made with a linguistic notion, namely Bernstein’s distinction between restricted and elaborated codes – which is also hardly illuminating because this sociolinguistic distinction of class-based language use has little to do with the more specific or generic types of master frames, such as that between that of Civil Rights on the one hand or the Nuclear Freeze frame on the other hand. Incidentally, Nuclear Freeze is rather a frame than a master frame, because it is only used by the Nuclear Freeze movement, and hence hardly generic. In this theory Master Frames are more flexible, more inclusive and therefore have more appeal and influence, because they have greater ‘narrative fidelity’ because their ideas are more central, and ‘ring true’ with extant beliefs, myths and folktales. Despite these comparisons and other characterizations, we still don’t know what, for instance, the Civil Rights master frame looks like – and how it is related to general attitudes, ideologies, or values.
Finally, the theory of master frames is applied to the notion of cycles of protest, by a series of propositions. For instance, a new innovative master frame may stimulate the activity of an SMO, as was the case of the Nuclear Freeze master frame for the peace movement, or if a master frame fails, this may explain why mass mobilization is not successful. Movements early in a cycle of protest are often progenitors of master frames. The concrete illustrations associated with these and other proposals about the role of master frames are enlightening, but since no explicit theory of their nature and structure are provided, the proposals remain general ideas for research, but do not prove how master frames are related to cycles of protest and their more specific frames, goals, values, ideologies or activities.
In a later encyclopedia article, Benford (2013) explains that the notion of master frame was coined in order to account for cycles of protest in the absence of a Political Opportunity Structure (POS): ‘A master frame refers to a generic type of collective action frame that is wider in scope and influence than run-of-the-mill social movement frames’.
Obviously, if the notion of frame is vague, the notion of master frame must be even vaguer, because it must apply to several social movements and sociopolitical situations. It may thus be defined in terms of overall norms or values (such as injustice), ideologies (imperialism) or overall goals (market choice).
There are many studies of master frames. Some of them, and the actual notions studied, are summarized in Table 2, which again shows that in these studies the notion of (mater) frame is analytically superfluous.
Studies of master-frames and the framed notions used.
Similarly, in their review also Hall and White (2008) critically show how master frames have been informally characterized (‘defined’ would suggest precise definitions) with such variable notions as meta-cultural themes of economic growth (Skillington, 1997), a society’s dominant cultural elements (McCaffrey and Keys, 2000), or discursive field (Fiss and Hirsch, 2005), scientific rationality (Mercer, 2002; Roth et al., 2003), social justice (Edwards, 2006), and local knowledge (Brown, 1992; Harrison et al., 1998). These various characterizations again suggest that frames and hence also master frames are not a specific conceptual or discursive structure, but a very general label for many very different properties of discourse and cognition, such as themes, ideologies, norms, values, goals or even contexts or opportunity structures (such as discursive fields) of SMs.
Other notions of frame analysis
Similar critical analysis may be provided for the many other frame-related notions in SM research, such as frame-resonance, framing effects, counter-framing, among others (see, e.g. Babb, 1997; Berbrier, 1998a; Connolly, 2005; Holland, 2014; Ivakhiv, 2010; Kubal, 1998; McCammon’s, 2009; McCammon et al.’s, 2004; McCammon’s, 2012; Schrock et al., 2004). As explained above for frames in general, also a study of ‘frame resonance’ would require a detailed cognitive analysis comparing various kinds of beliefs (models, attitudes, ideologies, values, goals) of SMs with those of adherents. Despite the interesting sociological contributions of these studies, this interest is not due to the use of the notion of frame, but rather to the use of the other notions (beliefs, etc.) dealt with in the analyses.
Frames and related notions
We have seen that framing studies actually analyze many different kinds of discourse structure and mental representations, phenomena for which there are more explicit theoretical frameworks. Fortunately, some of these other notions have been dealt with in other SM studies, as is the case for
We have seen that many studies of ‘frames’ presuppose a
One of the major criticisms formulated above is that most framing studies use
Yet, the specific framing papers of our corpus hardly refer to, or apply discourse analytical studies, not even of (social psychological) ‘discursive psychology’ (mentioned in two papers). Prominent sociopolitical discourse analysts, writing since the 1980s, such as Norman Fairclough and Ruth Wodak, whose work is directly relevant for the study various social movements, have only been cited in a handful of papers.
Hall and White (2008), quite critical of frame analysis, most frequently (N = 15) use the notion of ‘discourse analysis’ in their paper analyzing the testimony given by 115 witnesses appearing before U.S. Congressional committees in debates on salmon recovery policy. However, instead of detailed discourse analysis, they again use the traditional method of ‘coding’ of a ‘wide range of potential frames, including scientific, democratic, procedural, ideological, economic, and injustice frames’. They focus on specific discursive strategies ‘used to accomplish the core tasks of diagnostic and prognostic framing and to establish the credibility and salience of claims’, instead of actually analyzing the discourse structures of these ‘frames’ so as to show how speakers in U.S. congress use these ‘frames’.
Unfortunately, many of the SM studies purporting to use ‘discourse analysis’ still make very limited use of the many theoretical notions and methods of discourse studies carried out before the publication of their own articles.
Although more bibliographical research would be needed, our data suggest that most framing studies barely refer to studies and disciplines that could provide more explicit insight into the precise nature of (cognitive) frames and their expression in discourse.
Frame analysis in media studies
Articles on framing in media studies are not reviewed here and need separate critical review (among a vast number of studies, see, e.g. Entman, 2004; Iyengar, 1991; Johnson-Cartee, 2005). But the criticism formulated in this paper generally also applies to the analysis of frames in media research. For instance, Entman defined the notion of frame as follows:
To frame is to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation for the item described (Entman, 1993: 52).
The problem of this definition is that this (cognitive!!!) ‘framing’ aspect of discourse production does not result in the kind of (discursive) frames scholars aim to find and analyze in media text or talk. Salience (as a cognitive and discursive notion) can be expressed by a large variety of discourse structures, such as semantic macrostructures (topics), themes and their expression in headlines, leads, thematic sentences, keywords, summaries, sequential order, repetition, (lexical) frequency, metaphor, typography, size, photographs, schemas, position on the page or in the newspaper, as well as many other (semiotic) properties of online news and media messages. Thus, salience may be distributed in many ways in discourse (Chiarcos et al., 2011).
Perhaps the (media) discourse properties most closely corresponding to a (limited) definition of ‘frame’ are the semantic structures that define the ‘global meaning’ or ‘topics’ of discourse. These conceptual structures are typically expressed in headlines, leads, summaries and conclusions. It has empirically been shown that this ‘global’ information is first planned in production and best recalled in reproduction (Van Dijk, 1988; Van Dijk and Kintsch, 1983). This notion also comes close to the familiar sociological concept of ‘defining the situation’ we mentioned above. Cognitive analysis of such a concept, for example, in terms of (the top levels of) mental models, or the frame structures of concepts studied in frame semantics, both related to underlying knowledge structures, might provide more explicit definitions and applications of frames in the study of social movement discourse.
Conclusions
Our detailed critical review of many frame and framing studies in social movement research confirms and elaborates many of the observations and conclusions made above. Summarizing these conclusions, we found the following:
The main contribution of framing studies in SM research is the aim to focus on the
Despite this cognitive and discursive turn in SM research,
The methodological application of
Since the notion of frame, as used in SM research, has hardly been made explicit in cognitive or in discursive terms, and hence also cannot be applied in empirical research, it is recommended that
Finally, we can only speculate why the notion of frame became so popular in the study of media and social movements, especially since other, more specific notions were available to analyze discourse, cognition and (inter) action. My best bet is that it seemed to bring an attractive analytical method that could be applied to any data, without needing expertise of language, discourse or cognition. Indeed, like the ubiquitous metaphor of the picture frame.
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Author biography
.