Theorizing communication. The production of theory in communication journals

ABSTRACT

If communication research is in a ‘theory crisis’ or if there is at least an unused potential for theorizing, one of the ways of addressing such challenges is to analyze actual strategies of theorizing. This is not only relevant for practical and educational purposes but can also be a starting point for a science and technology studies’ perspective on communication research. This article presents a first qualitative study of structures of theoretical frameworks and of strategies of theorizing employed in articles in major international communication journals, focusing not only on the theoretical sections proper but on all aspects of theoretical argumentation. The analysis identifies a multitude of strategies of conceptualization and recombination, perspective shifting, and critical analysis.

KEYWORDS:

• Theorizing
• theory building
• qualitative analysis
• communication journals
• science and technology studies

How do you develop theories? Well, you may respond, you read previous literature, and then – if you are lucky – you come up with ideas (which you may play with, refine, discuss, and ultimately write up). On the contrary, asked how to collect certain types of data or analyze it, you may point to methodological literature, explain certain methods in great detail, remember courses on methodology you may have taken, etc. Certainly, while some of you seem to be experts in theory development, others struggle with the theory sections of their manuscripts, maybe compiling them last, pieced together from excerpts of the literature and a number of ad hoc arguments. A few of you may have read books on how to build theories (some of which are cited below) or be familiar with grounded theory. Both approaches are applicable in some cases, but less so in others. They mainly work when it comes to causal explanations as the main topic of the literature on theory production or to theorizing from qualitative data as the domain of grounded theory.

In this study, I take a broader look at the construction of theories in communication research, on their structure, and the operations of producing them, not only by discussing different options theoretically, but by analyzing empirically what methods of theorizing¹ (establishing, modifying, or productively applying theories) have actually been used in the literature. It is thus a study not of the content and implications of theories (their domain and conditions of application, the concepts and relations between them, their claims and underlying assumptions, their affiliation with paradigms, biases etc.). It is a study of the form of theories and, most importantly, of the practice of theorizing as opposed to finished theories.

Thus, the research question underlying the subsequent analysis will be:

What types and structural forms of theories and, more importantly, what strategies of theorizing (in the sense of producing theories) can be found in existing literature—in the present case—in articles in communication journals?

Why study theorizing in communication research?

To the best of my knowledge, this is the first study of theorizing proper (as opposed to finished theories and established paradigms) based on an analysis of contributions in major international communication journals. Findings on actual theory production can have different functions:

1. If many struggle with theory construction, it would be a good idea to teach not only theories but theorizing. To date, there is only limited and often selective literature on teaching theorizing (e.g. Gigerenzer, 2010; Silver, 2019; Swedberg, 2014b, 2014c, ch. 7, 2016). When theorizing is included in the curricula or taught in workshops in a given discipline, we may emphasize the strategies of theorizing that have been identified as most relevant by an empirical study in that field. They may also be most likely to inspire further theorizing and should therefore be identified empirically.

2. However, teaching and inspiration by existing examples also requires a critical reflection of current strategies of theorizing with regard to biases, problematic presuppositions and implications, and unused potentials. Collecting and categorizing prevailing strategies is a precondition for such a critical assessment, for example on the role of theory in the reproduction of racism, sexism, colonialism etc.

3. More broadly, several disciplines have diagnosed crises that call the value of their research output into question. Researchers observed a crisis of reliability that they think can be solved methodologically (e.g. by means of replication or analyses of publication bias) or institutionally (e.g. by means of open science or preregistration). However, for example, some psychologists have diagnosed a ‘theory crisis’ they consider even worse than the replication crisis (Eronen & Bringmann, 2021) or that they see as a neglected factor contributing to it (Oberauer & Lewandowsky, 2019). Furthermore, communication researchers have also long criticized the lack of theory production and theoretical syntheses and themselves declared a ‘crisis of theory’ (Shoemaker, 1993, p. 147) or an ‘interpretation crisis’ (Krämer, 2022) – a crisis in which we do not know what our findings mean, in which their validity is in question and their implications are not systematically reflected due to a lack of systematic theorizing. Even if we do not completely agree what is at the heart of the theory crisis (is it the lack of theoretical productivity or a problem of confusing overproduction and fragmentation, a lack of formalization or of well-founded interpretations and critical reflection? See, e.g. Krämer, 2022; Berger, 1991; Craig, 1993; Muthukrishna & Henrich, 2019) or even that there is a crisis, theory production can always be improved.

   In sociology, a need for more systematic theorizing has also been felt, which led to growing literature on types and functions of theory and, most notably, on theorizing, which will briefly be reviewed in the following section. The present study adds to the literature on theorizing in a particular way: by investigating theorizing itself, empirically or interpretatively. It may thus help to contribute to overcoming a possible theory crisis, at least in communication research, or at least to improve theorizing in different ways by practicing it more systematically, following strategies that have been identified in existing literature.

4. Finally, the study can be seen as a starting point for the study of scientific practice in communication research, in line with a science and technology studies (STS) and/or sociology of scientific knowledge approach. Such a perspective, which is only in its infancy (Kuznetsov, 2019), can open up new potentials for self-reflection in the social sciences and produce new insights by comparing the results from the studies on the social-scientific field to the findings on other disciplines.

If this study deals with ‘theorizing communication’ and analyses the strategies of theorizing in ‘major communication journals,’ a short reflection on the meaning of ‘communication’ and the relevance of such a study in communication research specifically is in order. It would be contradictory to provide a substantial definition of communication because such a commitment would preempt certain findings (as such a definition could be biased in favor of or against certain schools, whether social-scientific in the narrow sense, rhetorical, critical etc. with their predominant strategies of theorizing). Therefore, an institutional definition seems more appropriate, focusing on communication research as a discipline. In the present case, communication research is represented by what is published in the journals of a major scholarly association. Of course, such a definition introduces its own biases. The analysis is then biased towards the theoretical approaches that are accepted at the more powerful pole of the field of a differentiated community of practice. However, as this study does not analyze the content or perspective but the formal strategies of construction of theories, the biases may be somewhat subtler compared to the wholesale exclusion of certain substantial approaches, as will be discussed in more detail in the conclusion.

The need for a study of theorizing in the specific field of communication research can be justified based on three of main factors working against thorough theorizing in the discipline:

1. its inter-disciplinary character, leading to the ‘import’ of theories (Berger, 1991; Hagen et al., 2015) with sometimes only superficial adaptions to one’s object of research (instead of elaborate theorizing that can of course always be based on previous theories from other disciplines),

2. the impression that our objects of research are constantly and quickly changing, leading to descriptive studies of the new or of the developments (instead of relying on theoretical frameworks that are sufficiently general, comprehensive, and adaptive as to analyze new phenomena and/or the processes of change, sometimes also relativizing its presumed speed or impact) (cf. Beck, 2003; Stanyer & Mihelj, 2016), and

3. the heteronomy of the field (Park, 2014, p. 119): It is dependent on various practical fields, with structures, resources, and expectations that demand and rewards teaching of practical skills or applicable knowledge (Berger, 1991) instead of detached or critical, often abstract theorizing (although this is not necessarily a tradeoff, it is often perceived such). And, like other disciplines, parts of the field are dominated by indicators of research output that reward third-party funding (that is mostly granted for empirical and sometimes applied research instead of theoretical projects) and numerous publications based on small-scale and rarely theory-heavy studies.

Therefore, a study of existing strategies of theorizing in the field of communication itself can be particularly informative to those who aim to address the shortcomings in practices of theorizing, even if structural constraints remain in place.

Theory and types of theories

This section aims to clarify the understanding of theory that is at the basis of the present study, before the next section can elaborate on the shift of interest from theory to theorizing that underlies the analysis.

There is no consensus as to what exactly counts as theory or whether the word ‘theory’ can be said to refer to a unitary category of entities (Abend, 2008). It is therefore important to acknowledge that: (1) there is a continuum of types of statements usually made in the scientific literature, some of which would be classified more readily as ‘theoretical’ than others; and (2) to answer the question, ‘What is the meaning of “theory”?’ not with a single definition but with a list of different possible meanings.

Alexander (1982, pp. 2–3) postulates a spectrum of scientific statements ranging from the metaphysical to the empirical or from general presuppositions to specific observations (via, e.g. concepts or methodological assumptions). Where one draws the line between the theoretical and the empirical (and/or the theoretical and the meta-theoretical) is somewhat arbitrary. In the following, I am interested in statements that are certainly more on the generalized and speculative or argumentative side than on the side of systematically gathered new and specific information or data, while acknowledging that, for example, what is usually classified as ‘findings’ or ‘practical implications’ is always theory-laden and can thus provide interesting material for the study of theorizing.

Different meanings of ‘theory’ may refer to different sub-categories within the same category (that may in principle be delineated by one definition) or different – but of course not unrelated – categories that have come to be designated with the word ‘theory’ but that do not necessarily overlap much. Abend (2008) has proposed a list of different meanings of theory in the social sciences that may be helpful when classifying what can be found in the corpus of the present study (see also Büttner, 2021; Merton, 1945, for other classifications):

1. ‘a general proposition, or logically-connected system of general propositions, which establishes a relationship between two or more variables’ (Abend, 2008, p. 177), and based on this;

2. an explanation of a specific phenomenon logically deduced from 1.;

3. interpretive theory;

4. the exegesis of the works of ‘classical’ authors;

5. a ‘Weltanschauung,’ i.e. a conceptual, ontological, and/or epistemological framework to clarify the nature of the social world and one’s perspective on it;

6. normative theory of society; and

7. solutions to specific ‘philosophical,’ fundamental problems in the social sciences, which are not whole ‘Weltanschauungen’ in the sense of 5.

For the purpose of this study, ‘theory’ is then defined broadly to include:

• all statements and systems thereof that only stop short of specific empirical statements and findings, and that thus remain at least somewhat generalized, speculative, or argumentative, including

• the main theoretical framework presented in a publication (that can fall into to any of the above categories, but most often also with other parts belonging to other categories), and

• other theoretical, i.e. explanatory, interpretive, critical etc., claims and reflections in the broadest sense and the arguments used to establish them (e.g. epistemological and methodological assumptions or justifications of the relevance of a theory or study) that are not immediately connected to the main theoretical framework of a publication, including in parts of manuscripts other than the usual ‘theory section.’

This broad and rather enumerative than intensional definition was also chosen in order not to exclude specific types and functions of theories or paradigms of research. Furthermore, researchers focusing on one of them may always profit from strategies of theory construction that would, at first glance, mostly fit other types and paradigms (for example, someone establishing causal theories may well use some of the strategies most often adopted for the production of interpretive theories identified in this study or elsewhere).

Similarly, analyses of theorizing as the present one do not have to limit themselves to theories of a specific range. For example, communication research is said to be dominated by ‘theories of the middle range’ in a loosely Mertonian sense (see Merton, 1968; Schmid, 2010, on the ambiguities of the concept) instead of producing what, starting with Mills (1959), is often called ‘grand theory.’ However, this is both correct and incorrect. What is presented in publications in communication research are often applications of middle-range theories (insofar, the observation is correct), but as the result of this application, what is produced in terms of theory is more specific than middle-rang theories. Other authors present more specific ad-hoc conceptualizations and hypotheses that are not derived from any more general theory, whether middle-rang or ‘grand theory.’ Finally, some authors apply (but only rarely develop) more general or wide-ranging theories that would be classified under the equally ambiguous term ‘grand theory.’ It would therefore be unwise to limit a study of theorizing to a – difficult to define – range or level of abstraction. It is also unnecessary because certain strategies of theories apply to different ranges of theory. For example, Talcott Parsons, who used to be admired and disparaged as the incarnation of the ‘grand theorist’ when the term was coined, made excessive use of cross-tables of conceptual distinctions – a tool that can as easily be applied to generate the most specific ad-hoc hypotheses or typologies.

From time to time, researchers have proposed classifications and descriptions of general approaches, schools or paradigms in communication research (e.g. Craig, 1999; Lang, 2013; Rosengren, 1983) or quantified their use, the use of specific theories, the relevance of theorists, or references to criteria for the assessment of theory (e.g. Beniger, 1990; Bryant & Miron, 2004; DeAndrea & Holbert, 2017; Fink & Gantz, 1996; Kamhawi & Weaver, 2003; Potter, 2014; Potter et al., 1993).

Although it goes beyond this line of research, the present study will also typologize theories, but specifically focusing on their structure – their elements (the concepts and how researchers actually deal with them), the relations between those elements (e.g. x causes y; x can be interpreted as y, etc.), and the resulting claims and functions of a theory (i.e. to explain, to interpret, or to critique, Potter et al., 1993).

From theory to theorizing

In a turn that is quite often used when theorizing, the present study then shifts focus from something static or structural, from objects or products – theories – to something dynamic or performative, to practice: theorizing as the process of producing theory. Below, I will further elaborate on the meaning of theorizing and what aspects of it the present study focuses on.

I am only aware of one study specifically analyzing theorizing in the field of communication (Hagen et al., 2015), which focuses on German-speaking journals. It quantifies, for example, the share of data-driven theorizing and of theory production that works rather from scratch or with existing theories. Otherwise, the growing literature on theorizing has hardly been cited in communication research (however, as an exception, Shoemaker et al., 2004, have even directly contributed to it).

Slater and Gleason (2012) distinguish and also quantify different ways of ‘contributing to theory’ in communication journals (see also DeAndrea & Holbert, 2017). However, they only include quantitative studies and use pre-defined categories at different levels of abstraction that would partly count as strategies or operations of theorizing proper and partly as to the ways finished theories or studies contribute to ‘generalizable knowledge’ (Slater & Gleason, 2012, p. 215). For example, their categories include, ‘Adding a theoretically relevant independent variable’ (which I would consider an operation of theory production proper, i.e. creating or modifying theories), ‘critiquing and testing key theoretical assumptions’ (very general functions or logics of scholarship that do not so much modify or establish theories as to judge them in the light of arguments and/or data), or ‘asserting a causal process because a structural equation model of the proposed process provides good fit’ (a very specific type of contribution to knowledge based on a specific type of model and analytical tool).

Perhaps most prominently, Swedberg (2014c) has described theorizing as a practice that can be taught and learned, and that leads to the increasing acquisition of skills that allow theoreticians to leave the scholastic rules behind. Theorizing should not be understood as creatio ex nihilo by the demigod genial theorist but as an activity that draws on pre-existing elements, and the choice of this ‘raw material’ is an important decision that enables and restricts all further activities of theorizing. Theorizing thus does not only require practice but also a familiarity with existing concepts, theories, or theoretical mechanisms (Swedberg, 2014c, ch. 8) before applying different methods or strategies.

For the purpose of the present study and in line with the broad definition of ‘theory,’ theorizing also includes the application of existing theories or concepts to new phenomena (as opposed to its transformation into designs and instruments of data collection). This is a generative act because it involves operations of matching or deduction and creates an understanding of a phenomenon that has previously not been analyzed in terms of the framework applied to it – a new theory of the phenomenon, at least in the above broad sense of ‘theory.’

In principle, theorizing from data also falls into the scope of the study. However, as far as articles follow usual methodologies of coding, paraphrasing, comparison, or conceptualization as described in textbooks on qualitative data analysis, such practices will mostly be left aside in the results section. At present, their analysis would not add as much to our knowledge on theorizing as an analysis of the strategies of theory production that go beyond such textbook methodology (even if it is always worthwhile to analyze actual scholarly practice as opposed to its representation in the literature, and future studies should therefore study practices of qualitative analysis more closely from an STS perspective). However, more elaborate strategies of conceptualization and of combining concepts will be elaborated on below even if they appear in results sections, not only in the more ‘speculative’ or interpretive and argumentative sections of papers.

It is impossible to catalog each general methodology and specific method mentioned in the previous literature on theorizing in this article, but I can only mention what I consider to be the main approaches. This ad hoc typology based on my review of the literature should then be complemented by an empirical typology as will be established in the present study.

Some authors have suggested different ways of thinking intuitively and abductively once an interesting phenomenon has been identified, named, and conceptualized (Swedberg, 2012) or described general ways of dealing with theory such as transferring or combining theories (Byron & Thatcher, 2016). Others have identified general templates for theories, i.e. general theoretical orientations that propose fundamental causal mechanisms for the explanation of phenomena (e.g. Bell, 2009; Elster, 2007; Rueschemeyer, 2009; Stinchcombe, 1986). Some systematically go through the possible combinations of a set of two, three, and more variables or describe techniques of creative thinking to bring together elements of theories and to construct explanations (Shoemaker et al., 2004). Still others have focused on the possible steps of formalizing theories once a proto-theory has been developed, for example by analogy to other phenomena (Borsboom et al., 2021; see also Jaccard & Jacoby, 2019, on mathematical modeling and simulation).

While each contribution on theorizing has its strengths as it focuses on certain useful strategies, we still lack a systematic compilation – in particular of the strategies researchers in our own field have actually applied for their publications and thus obviously found to be fruitful. Those strategies will be collected in the present study, which should then be complemented by other typologies established from other sources (such as interviews with researchers and types of literature other than journal articles) to broaden the picture and identify unused potentials for theorizing.

As some have criticized (e.g. Büttner, 2021; Krause, 2016), the literature on theorizing has mainly covered the development of causal explanations. However, many researchers wish to conduct theoretically informed interpretive research or adopt a critical perspective on their research topics. These and other researchers who (momentarily) do not focus on causal explanations may also be interested in strategies of theorizing that are appropriate for their task. Therefore, the present study will cover the development of a wide variety of types of theories across different levels of abstraction and generality. This will also include minor theoretical aspects in all parts of journal articles, such as normative conclusions in a contribution mainly interested in causal analysis. Thus, the present study goes beyond monolithic classifications of theories or approaches and aims to shed light on the many places where theory in the broader sense comes into play and on the overall theoretical frameworks that go beyond a core of causal hypotheses or concepts informing an interpretive analysis.

Theorizing as practice and its analysis

Theorizing is a practice that can be studied for at least three purposes, as noted above: to learn and teach how to theorize, to learn more about scientific practice, and to critically reflect on this practice. Practices, in turn, can be studied in three ways: by means of observation, self-reports, and based on their traces or results. The present study relies on the last option for several reasons.

From an educational perspective, it seems useful to draw on a sufficient number of cases instead of starting, for example, with a case study of theorizing in one or a few research projects. An analysis of several dozen articles is therefore preferable. Although it may certainly also be relevant to interview scholars about their strategies of theorizing, it is questionable whether this would be overwhelmingly more fruitful than an analysis of their articles if they fail to remember their methods of theorizing or to verbalize them precisely. However, this avenue should also definitely be pursued to complement this study.

Of course, most publications do not come with a ‘methods section’ describing their methods of theorizing. Going beyond a quantitative survey of strategies described explicitly in publications (as done by Hagen et al., 2015, or to a certain degree by Slater & Gleason, 2012), a qualitative study can infer and explicate methods from cues in the material. Furthermore, it may be wise to demonstrate how to ‘reverse engineer’ the theorizing of others from their works (see Swedberg, 2012, for a similar suggestion, and Krämer, 2015, for an example). The readers of the present study can then apply this methodology to other texts.

An analysis of theorizing in communication research can also contribute to an emerging STS perspective on the social sciences. The few existing studies seem to have focused on data collection as the performative constitution of social reality as well as tacit knowledge in survey research (Law, 2009; Maynard & Schaeffer, 2000). Knorr-Cetina (2014) describes the practices and experiences of intuitive theorizing in way that is similar to ethnographic analyses in STS but relies on casual (self-) observations. This state of research leaves room for more systematic analyses of theorizing and overall research practices in the social sciences (see already, e.g. Krey’s, 2022, ethnographic analysis on writing in sociology). Such a research program would at best be informed by a previous careful interpretation of publications that reconstructs the tacit knowledge on strategies of theorizing, as in the present study (cf. also Martus & Spoerhase, 2022, for an analysis of research practices in the humanities based on published and unpublished written sources).

Although we can learn a lot from the cues in articles, we should be aware that their theoretical argumentation is, to a certain degree, constructed ex post and that the composition of an article is not simply an accurate representation of the mental, bodily, and material practices of theoretical thinking, discussion, note-taking, and drafting (see also Abend, 2006). In the future, publications in communication research, and the social sciences in general, may therefore also be studied more explicitly from a rhetoric of science perspective (Harris, 2018, 2020) – that does not, however, simply oppose ‘mere’ rhetoric to ‘actual’ research, and instead analyzes how different aspects of practices and texts establish knowledge.

Taken together, the above-mentioned analyses on data collection, future interview and ethnographic studies, and the present analysis may be a good starting point for an STS approach to communication research and the social sciences.

Method

This study is based on a qualitative reconstruction and typology of the fundamental type and structure of theories and the underlying strategies of theorizing in major communication journals (see Abend, 2006, for a very similar methodology and design in his analysis of understandings of theory in US and Mexican sociology). The sample includes articles from three journals published by the International Communication Association that represent different paradigms and schools of communication research. These journals publish different types of articles, either serving as a broad forum for the discipline or with a substantial amount of empirical and, often, quantitative studies (Journal of Communication), focusing on theory (Communication Theory), or with specific emphasis on qualitative and critical analyses (Communication, Culture and Critique). Together, these journals should cover a wider variety of topics and most of the established paradigms in the discipline.

As a starting point, all articles in the latest issue of the above journals at the time of the analysis (issue 1/2022) were included to get an idea of the current state of theorizing in the outlets. To avoid biases related to the decisions of single editors, the articles from a complete issue in the middle of the previous editor’s tenure were then added. After a preliminary analysis of these papers, further articles from the three most recent volumes were selected based on the likelihood of contributing new insights or challenging an emerging impression of theoretical saturation (the 65 articles in the final sample are listed in Appendix A).

The analysis proceeded along the following steps:

1. In all sections of the articles, theoretical statements were identified (including the main theoretical framework and other claims and reflections that were ‘theoretical’ according to the broad definition above – in this sense, there were no completely ‘atheoretical’ articles in the sample).

2. The main theoretical framework was shortly paraphrased, focusing on its type (e.g. causal, critical etc., following the above classification by Abend, 2008) and structure (e.g. a causal network or a model of a process with several steps), not the content of its concepts and claims.

3. All statements by the authors about how they established their theoretical framework and claims were identified and paraphrased (e.g. authors stating that they introduced a new moderator to an established causal relationship or that they viewed phenomenon X through the lens of theory Y). If the process of theorizing was not explicitly described or the description remained vague, the strategy of theory production was inferred from the structure of the theory as compared to its ‘material’ (previous theories, concepts etc.) and other information in the article (e.g. if the authors present a fourfold classification in a 2 by 2 matrix, it seems clear that this theoretical typology has been established by crossing two dichotomous distinctions).

4. This process led to a classification of types of theories and strategies of theorizing that was iteratively modified (differentiated, complemented, or simplified) by comparing the individual cases among each other (e.g. a fourfold matrix is similar to a coordinate system with four quadrants because both can be established by crossing two concepts) and to types of theories and strategies of theorizing mentioned in previous literature.

Results

Types of theories and materials for theories

Many of the main theoretical frameworks used and built in the articles in the sample fall into a number of relative clear-cut types with regard to their structure and function (more or less in line with the above-mentioned typologies, mainly nos. 1., 3., and 6. in Abend’s classification):

1. Causal theories establishing a relationship between factors and outcomes, often via several causal steps or including mediators and moderators;

2. conceptual frameworks or frameworks of interpretation that (a) establish (postulate or synthesize) or critically discuss definitions of concepts and/or (b) interpret phenomena in terms of a broader trend or certain concepts, categories, or theories (sometimes, such frameworks take the form of a more encompassing categorization, typology, or other structure, but mostly consist of loose combinations of concepts from different sources); and

3. normative theories and critical theories. This is a rather broad category, the different theories do not necessarily have much in common, and different strategies of critical analysis will be discussed below. These types of theories can nevertheless be grouped together based on the types of claims that they make: They are not (only) descriptive of reality, but evaluative with regard to problematic features and relationships, inherent contradictions, or in postulating desirable states of affairs. Other non-descriptive theories or elements of overall theoretical frameworks include directive statements, i.e. statements that ask or demand of others that they do something. This takes the form of programmatic statements, e.g. our discipline should turn to the question of the good life (Vorderer, 2016 in Appendix²), or simply: ‘further research should … ,’ or reminders, appeals, and exhortations to study things in a specific manner.³ Examples may include guidance to not adopt decolonial perspectives in a superficial way but realize the ‘decolonial possibilities yet unfulfilled’ (Na’puti & Cruz, 2022 in Appendix, p. 2); or conceptualize incidental exposure as dynamic, not static (Chen, Kim, & Chan, 2022 in Appendix), etc.

Everything in an article’s theoretical framework may not fall under same category. Instead, causal theories need conceptual frameworks and interpretive statements, except that causal theories often do not discuss their conceptual framework in much detail, simply providing definitions or leaving most of the conceptual basis implicit. The conceptual work is not what makes them causal theories. Interpretive theories, in turn, sometimes rely on causal assumptions – while often avoiding causal language, speaking for example about ‘implications,’ something being ‘informed by’ something else, etc., instead of ‘effects’ or ‘influences.’

Directive statements are rarely at the center of theories (although the contribution by Vorderer, 2016, is an example of a programmatic text). Descriptive theories are often accompanied by normative judgments in concluding reflections or to justify their relevance (however shortly they are mentioned – maybe just an expression of concern, optimism, etc.). Authors also justify theorizing by the demands of those they cite (‘X has called for … ’).

Theoretical frameworks or references may include functional analysis or analysis in terms of problem and solution (If x exists, what is its function, what problem does it solve? Or: If we have problem y, what could be the solution?). Other elements include methodological, epistemological, ontological, and statements on good scientific practice.

For example, to Riddle and Martins (2022 in Appendix), replication means to strictly follow an original methodology in an ‘apple-to-apple’ approach. Huskey et al. (2022 in Appendix) commit themselves to the idea that science is cumulative and best advanced by replication, providing critical, new information. However, according to the authors, this can mean to vary previous designs for less ambiguity – even if this means that in the case of unsuccessful replication, hypothesized causal relationships and methodological aspects of the study may have to be called into question. Authors also hint to certain ontological assumptions, for example, when Bock (2016 in Appendix, p. 18, citing Nick Couldry) states: ‘The world is not a text,’ and emphasizes embodied practice.

Although scholars of course write on methodological and practical matters of research and although their arguments have ontological implications, no theories that were exclusively of the ‘Weltanschauung’ or ‘philosophical’ type were found in the volumes of journals included in the analysis (Vorderer, 2016 in Appendix, comes closest to this type when he postulates three universal and basic needs as the basis for the good life and hints to metahistorical beliefs, such as that utopian ideas can turn into dystopias). Furthermore, no primarily exegetical, critical, or comparative articles on other authors’ works were published in the issues scanned for the analysis (although several authors review the history of certain concepts), and methodological and epistemological questions in the strictest sense were neither the focus but often briefly covered. This is, however, more a matter of differentiation within the field than of total neglect, as such texts are published elsewhere, even if rarely.

Before turning to the actual strategies of operations of theorizing in the next section, I will also shortly address the ‘material’ for theory production, the existing elements that are combined and transformed into the new theoretical frameworks or statements. The types of material include:

• existing theories and concepts, including commonly accepted distinctions and classifications, some of which do not (or no longer) belong to any one specific theoretical framework or that one would not classify as ‘theoretical’ in the strict sense but foundational to communication, social, or psychological theory (between the communicator, medium, message, and the audience/reception/effect; between more or less systematic, elaborate processing; between ethos, pathos, and logos; micro, meso, and macro level, etc.);

• empirical findings that are theory-laden but presented in a way that does not (yet) tie them to any specific theory;

• current public discourses, social trends, and much-discussed research topics (such as the rise of social media, polarization, or the assumed rise of disinformation, even if an article is not specifically about those very trends); and

• metaphors, symbols, stories, etc., from different cultural repertoires. For example, Schraedley and Dougherty (2022 in Appendix) identify a discourse that is to others like a ‘maestro’ in that it ‘helps conduct, or move and organize,’ or like ‘a powerful gear’ (also illustrated graphically by cog wheels).

Strategies and operations

Applying theories and concepts

In the articles included in the study, theorizing most often means to apply one or several theories or concepts, often in a rather syncretistic combination of materials such as previous theories, existing and/or new findings. However, in many cases, this generative process is not described in much detail. Often, leaving this process in the dark, authors only state that ‘X [the object of study] is/can/must be seen as Y [a theoretical category].’ However, for example Trillò et al. (2021 in Appendix, pp. 876–879) elaborate on what it means that they analyze ‘Instagram as a user-generated archive’ instead of simply stating that they do so. In several other publications, the role of the applied framework is expressed metaphorically or very abstractly instead of explaining what it means to apply it to the object of study. For example, the theory being applied is called a ‘lens,’ a ‘map,’ an ‘inspiration,’ a ‘background,’ or the authors mention being ‘indebted’ to someone. Such metaphors do not provide any further information on how exactly the authors proceeded to connect their object of study or discussion to the theories or concepts, or how they justify the application.

In general, authors almost never explicate the process of ‘application’ by relating the object and the general perspective in a point-to-point matching of elements and relations on both sides. Alternative frameworks are rarely discussed and reasons why one has chosen are rarely provided, but certain relations are justified based on a number of recurring types of arguments (see below). In addition to the deductive application of closed, overall theories or single concepts, authors can also enter a flow of associations, such as Scheible (2020 in Appendix) who establishes ever new ‘accidental allegories’ between the series Tiger King and society as a whole during the Covid pandemic.

Other theoretical operations are also sometimes described by metaphors, such as theories being ‘knitted together’ or ‘blended,’ brought into a ‘dialogue’ or ‘conversation.’ This again points to the necessity of explicating strategies of theorizing, which are not made entirely transparent by the original authors.

Operations of distinctions and recombination

A next set of operations deals with distinctions between two or more categories and with relationships between concepts. These operations establish or modify distinctions, and then connect distinctions or concepts by different types of relationships, or arrange them in different types of patterns.

In addition to the mere introduction of concepts or distinctions, different criteria or strategies of differentiation can be compared and discussed in an article, such as Baden and Sharon (2021 in Appendix) weighing different possibilities of distinguishing conspiracies proper from theories about conspiracy. Distinctions can be questioned or complicated, such as some authors arguing that online and offline phenomena can either no longer be distinguished or remain distinct but are strongly interconnected. Concepts can be extended or narrowed down.

Two distinctions can be combined into a fourfold table (or two continua into a coordinate system, resulting in four quadrants, e.g. Bighash et al., 2018 in Appendix). New elements can be built into existing concepts. For example, Graves and Spencer (2022 in Appendix) replaced the binary conception of truth contained in the concept of gaslighting by including uncertain, unverifiable, or probabilistic claims and a different overall conception of truth.

Apart from the relation between a theory or concept and a phenomenon they are applied to, another type of relationship that is frequently established is causal instead of (ana)logical. Operations based on causal chains are at the center of much of the previous literature on theorizing (and were already quantified by Slater & Gleason, 2012), and can be summarized very shortly. Authors prolong causal chains by adding new steps, they fuse established causal chains (with a dependent variable becoming an independent one in the next causal relationship), and add moderators and/or mediators. More complex operations include establishing feedback loops (e.g. Reveilhac & Morselli, 2022 in Appendix) or postulating spiral processes (e.g. Vandenbosch & Eggermont, 2021 in Appendix) that add an inverted or circular causality. Processes can also be ‘multiplied’ by describing the behavior of two or more entities in two parallel threads of a causal model and their interaction. For example, Grizzard et al. (2020 in Appendix) model the interdependent affective dispositions toward two characters in a narrative, establishing a symmetrical, ‘mirrored’ path model based on previous models, with causal chains for each character and adding interconnections between the dispositions toward the actors.

To open up new potential for operations, a theory or a set of relations can be mapped onto a given form and then operations specific to that form can be applied. For example, causal relationships can be described with a path model, then adding, deleting, or inverting paths. Russell and Reimer (2018 in Appendix) suggest conceptualizing mental models as semantic networks, so that measures from network analysis, such as centrality can be integrated into a theory of the processing of argumentative messages. Authors also visualize the elaboration of a model as such, for example, Kaun and Stiersted (2022 in Appendix) who first present a model with placeholders and then replace them with notions related to their topic. Grizzard et al. (2020 in Appendix) visualize successively more complex models based on balance theory, step-wise adding new elements.

Justifying relations

To elaborate and justify matches, such as between a general concept and more specific phenomena or between a cause and an effect, a number of modes of reasoning are used. Of course, first, logical deduction can establish a conclusion: If A leads to B and B leads to C (according to theory or previous findings), A should lead to C (as per the assumed transitivity of causality). Or if x belongs to the more general category of X due to its properties, which make it an X, and something has been argued to be true of X in general, then it should be true of x in particular. For example, if something is an action, theories of action apply, so if the theory of reasoned action is effective in ‘predicting human behavior across conditions and contexts,’ it can be used to predict specific (voluntary) behavior such as the adoption of innovations in interorganizational systems (Fu, Shumate, & Contractor, 2020 in Appendix, p. 500).

Second, many relationships are established by analogy. If proposition p is true of phenomenon x (e.g. x causes y; x has the property z; x can be seen as a case of X), then p will probably also be true of similar phenomenon x’ (e.g. x’, like x, will also cause y, etc.). For example, what works for web-based health interventions should also work for interventions based on texting (Gonzales, 2016 in Appendix).

Third, there is a remaining category of conclusions that can be vaguely described as reasoning per association. They are not justified in much detail but rest on an intuition that the concepts or phenomena brought together have something in common or are generally associated with one another. For example, a cause leads to an effect that is somehow similar; or two concepts with a certain semantic overlap or similar connotations will be related. For example, scanning (as opposed to information seeking) can be considered more ‘passive’ and peripheral processing also has connotations of passivity, so the two may be associated (Lewis, Rossmann, de Bruijn, & Martinez, 2022 in Appendix).

Strategies of perspective shifting

Other strategies of theorizing do not recombine existing elements but shift perspectives, leading to another view of the object of study. Some authors refer to ‘turns,’ such as the discursive or material turn (e.g. Hroch & Carpentier, 2021 in Appendix), that imply a new perspective. However, the strategies of shifting need to be explicated in more detail.

The first is a shift of focus that can simply mean to question and break away from dominant research interests and approaches and move toward others, for example, from certain parts of a causal chain to other steps, to other parts of the communicative process (e.g. from communicators to recipients), to different social groups (e.g. from dominant to marginalized ones) or to other levels of society. For example, according to Kaun & Stiernstedt (2022 in Appendix, p. 72), ‘Sheila Jasanoff and others have emphasized the importance of sociotechnical imaginaries for the nation as a project, but that they can also be situated in other collectives such as companies or social movements’ (emphasis added).

Beyond such shifts of attention, as important as they may be, more complex ones are possible. One option is to move from the analysis of a phenomenon to the study of meta-discourses, which may or may not imply that researchers remain agnostic about the existence of the phenomenon or the adequacy of discourses about it. For example, Abetz & Moore (2018 in Appendix) move from the analysis of the ‘mommy wars’ about parenting ideologies to discourses about the mommy wars, to how bloggers make sense of them. Or Bratich (2020 in Appendix) analyzes ‘fake news’ as the topic of a moral panic instead of studying, for example, the spread or effects of disinformation.

In a similar, potentially de-essentializing move, studies can take a constructivist or performative turn, analyzing how something is made sense of, its meaning is being negotiated or fought over, or how something is performatively constituted. For example, Trillò et al. (2021 in Appendix) do not consider values as something that we simply ‘have’ or share but investigate how their meaning is socially constructed through visualizations on Instagram and show how they are defined in a specific way – mostly apolitically, centered on the self and on aestheticized consumerism.

Another operation that somehow ‘dissolves’ fixed entities or states is to consider something as dynamic, to show how it manifests differently at different times, or in what sense it is a process instead of, for example, a stable entity or property. Conversely, the seemingly dysfunctional, inefficient, instable, and de-stabilizing can be re-interpreted as functional or enduring, making the phenomenon more ‘solid’ instead of dissolving it.

Instead of treating online groups as given and closed entities, Bighash et al. (2018 in Appendix) show how they are dynamic systems constituted by asking questions and providing information that can also dissolve if no further information is requested. Conversely, question-askers should then not be considered as ‘useless’ free-riders but make an important contribution to the constitution or perpetuation of such open, dynamic groups.

A final group of strategies based on different perspectives establishes contradictions. Contradicting findings or arguments can be used to stage a collision of perspectives that is either resolved by arguments or shifted to the empirical level by establishing alternatives hypotheses (e.g. Chen, Kim, & Chan, 2022 in Appendix; Huskey et al., 2022 in Appendix; Graves, Nyhan, & Reifler, 2016 in Appendix). Other strategies establish paradoxes, point to unintended consequences, or give their object a counter-intuitive form. For example, Korstenbroek (2022 in Appendix) argues that both the inclusion and exclusion of right-wing voices from the public sphere lead to undesirable consequences. Or websites for the recruitment of fixers are created with the intention to enhance their findability and status but perpetuate inequalities (Palmer, 2022 in Appendix).

Critical strategies

Usually, the interpretive and critical contributions in the sample do not provide complex theoretical frameworks. Rather, they mostly limit themselves to a number of concepts or labels loosely connected among each other and to the object of analysis (e.g. merely pointing out that they see X as Y, see X from a Foucauldian, Bourdieusian … perspective, etc.). However, more importantly, they employ a number of strategies that establish a critical perspective, which allows to challenge the status quo.

The first is to diagnose and question inequality, mostly with regard to an implicit standard of equality, restorative justice, or emancipatory potential (an operation of shift of attention and/or of comparison, performed with a normative stance). This is done, for example, by ‘highlighting and contesting historic asymmetries of power around race, gender, sexuality, class and ableism’ (Cepeda, 2021 in Appendix, p. 670), pointing to consequences of a phenomenon for marginalized and vulnerable groups or to bias and discrimination, or by centering on groups whose perspective is otherwise erased, such as indigenous people (Na’puti & Cruz, 2022 in Appendix).

In particular, such a perspective can be opposed to the appearance of equality, for example when a political process is ‘theoretically open to anyone but in practice favors larger, better-resourced organizations’ (Shepherd, 2018 in Appendix, p. 239), or something can be shown to function in favor of a political agenda and the privileged (such as conservative comedy ‘protecting the wealth and class privileges of America’s right-wing elite,’ Marx, 2022 in Appendix, p. 22).

Establishing contradictions has been identified as a strategy of theorizing above. Critical research opposes different explanations, constructs interesting counter-intuitive narratives, and similarly juxtaposes appearances and reality, revealing unintended consequences or explaining how a problem is not a matter of individual decisions but a structural problem. For example, Martin (2018 in Appendix, p. 294) shows to how the logics of casting ‘function to center ‘choice’ while concomitantly ensuring those choices tend to uphold existing ideologies,’ and, according to Ruberg (2020 in Appendix, p. 55), ‘the case of queer indie games demonstrates how the rhetoric of empathy, while often well-intended, promotes the appropriation and consumption of marginalized experiences.’ Critical studies also diagnose inescapable or remediable evaluative ambiguities and seeks to add complexity to the analysis of the seemingly unambiguous.

Critical analysis sometimes also goes a step further than measuring reality by certain established norms, instead challenging dominant conceptions and discourses by questioning the seemingly self-evident or universal, deconstructing essentialized representations of technology (Bourdon, 2020 in Appendix), politicizing the seemingly unpolitical, or de-naturalizing the seemingly natural. For example, Assaf and Bock (2022 in Appendix, p. 85) adopt a critical approach to textual analysis that identifies ‘ideological myth by seeking specific linguistic tropes, such as euphemisms, assumptions, and assertions of what is “natural” – particularly when what is “natural” justifies or reproduces existing social hierarchies.’

Even what is considered problematic by many may seem self-evident and self-perpetuating. Critical analysis can then instead reveal the ‘conditions of possibility’ of the problem. For example, Lawless and Cole (2021 in Appendix, p. 162) argue that ‘the spread of anti-intellectualism […] is a[n] […] effort that requires material resources such as hefty financial backing and a system for cheap, exploitative “journalistic” labor.’

Conclusion

This study would achieve a first goal if it were to sensitize readers to the diversity of the types of theory and theoretical statements, and where they can occur in a publication, and, more importantly, of the strategies of theorizing behind them. I hope to have demonstrated that theorizing can be analyzed in publications and consequently, be taught and criticized, and that its analysis may thus contribute to good research – an aim that cannot be realized by means of elaborate methodology and empirical practice alone, but that requires additional skills and reflection.

At the same time, this analysis will hopefully be one of the starting points for an STS (and/or rhetoric of science) perspective on communication research that emphasizes, for example, tacit knowledge and performativity. For example, together with other rhetorical devices and strategies of theorizing, the skilled use of metaphors for theorizing or to legitimize theoretical approaches is an example of a competence that is not taught explicitly and not readily acknowledged in a discipline that often considers neutral, precise, and prosaic language as most scientific. However, this mostly tacit knowledge on how to theorize and present theories (or on how to ‘tell stories,’ as Holbert & Park, 2020, phrase it) keeps research going and the results, if they are persuasive, can shape entire careers and fields.

Such a perspective also sheds light on the performative aspects of research. Individual scientific studies and whole fields constitute their objects of analysis through concepts and via ontological, epistemological, and methodological claims (and research does not only ‘make up’ groups or other social phenomena, but also affects them; see Hacking, 1986). Furthermore, approaches do not only convince through logical deductions and their compatibility with methodologically sound data, but also through the performative acts of appeals and exhortations: x must/must no longer be seen as y – otherwise, your research is old-fashioned/missing the point/ … ! x must be seen as dynamic – because who wants to be accused of static thinking?! x has to be analyzed critically – or who likes to be called uncritical?!

Of course, a textual analysis cannot describe the material practices and infrastructures of theorizing that an STS perspective would also be interested in but that remain to be analyzed in future studies based on other methods. For example, the syncretistic conceptual framework for the present study has mostly been elaborated looking out of the window of high-speed trains and collecting notes in a document on a laptop computer between reading papers. The present and similar findings should therefore be combined with data from other sources, such as interviews and ethnographic studies.

Such an STS perspective on specific practices of theorizing would complement other sociological approaches to the field of communication that emphasize the overall choices of, and struggles over approaches in relation to the distribution of resources in the field (e.g. Park, 2014, pp. 105–130). A closer analysis of practices will introduce additional levels of critical reflexivity in addition to the overall situatedness and power relations revealed by field analysis.

However, it seems useful to establish a categorization of possible practices of theorizing based on textual analyses such as the one presented above, in order to investigate them more closely or to be sensitized once they re-appear in the field. Such a sensitization seems useful given the assumption that strategies of theorizing are most often based on tacit knowledge and thus less easily identified than canonized or institutionalized research practices.

For example, if giving a concept a constructivist or semiotic turn is a strategy of theorizing according to the above analysis, this type of strategy can then be recognized in the field and we can further observe how it is enacted mentally, communicatively, and materially (imagine a whiteboard in an office on which someone wrote during a meeting: ‘DISINFO = FLOATING SIGNIFIER,’ and left it there as a constant reminder of a project’s approach to disinformation).

Taken together, these perspectives on theorizing can then also joins the vast literature in STS or the sociology, philosophy, and history of science that can be casually summarized as demonstrating that there is a difference between ‘textbook science’ (Kuhn, 1977, p. 360; i.e. science as it is supposed to be practiced according to textbooks or has worked according to simplistic or whiggish accounts of the history of science) and science as it is actually practiced.

I would also like to emphasize another goal of theorizing that I find highly important. Science communication is not only about conveying a simplified narrative about empirical findings but should also contribute to a dialogue about the challenges and crises of our time. While, of course, strategies of theorizing can create conceptual frameworks that can guide empirical research, similar practices can also be used to establish concepts and metaphors that can be introduced into public debate and shift perspectives in public discourses – which, in my impression, often has a stronger impact than empirical findings (for example, compare the seemingly self-evident concept of ‘filter bubbles’ with the often opposite empirical findings on users’ exposure to different views that have not yet been captured in a similarly appealing concept or metaphor). These concepts and metaphors should of course be consistent with research findings and can summarize and interpret them in fruitful ways.

If a wide range of strategies has been identified, is the status quo of theorizing satisfactory? Despite the encouraging diversity in the aggregate, a few critical but necessarily subjective remarks on specific features of practices will be necessary. A critique of the current practice of theorizing can concern different levels, some of which can be addressed based on the present analysis, and others would require additional data and/or an analysis of theories in terms of their substantial concepts and perspective on the social world.

For example, an analysis of theories’ perspective and relationship to their object (for example, their neglect of certain social perspectives in terms of race, gender, colonial history etc.) requires a different level of analysis. Similarly, an investigation of the social factors affecting theory production or the relationships of power around theories (who develops them, whether and where they can be published, whether and how they are cited, adopted, or canonized etc.) would mainly have to be based on further data on actors and institutions in the field. To my knowledge, we still lack systematic analyses demonstrating the consequences of structural inequality running from the production of theory, via publication and teaching, to citations and adaptations (although studies show inequalities in terms of thematic orientations, publications, and/or citations according to race, nationality, or gender; see, e.g. Chakravartty et al., 2018; Potthoff & Zimmermann, 2017; Trepte & Loths, 2020).

However, a number of critical conclusions can also be drawn based on the current analysis and on a reading of the underlying material. Using the above distinction between operations that recombine concepts and those that enable shifts of perspective, we may speculate that the former may carry a stronger risk of reproducing biases and relationships of power or appropriation related to canonical concepts (perpetuating a canon that reproduces, for example, colonialism and racism; see, e.g. Mohammed, 2022).

Apart from critical operations, those that shift perspectives can also de-center dominant knowledge, even if there is no inherent guarantee that this will always work in desirable ways. The distancing that is required for such operations may, on the contrary, absolve researchers from any substantial, but politically charged commitments if they decide to simply analyze the construction of something by ‘all sides.’ Furthermore, to distance oneself from dominant knowledge or pre-defined ‘social problems’ requires a certain degree of autonomy that can be difficult to maintain under strong hierarchies or financial dependence.

If, conversely, researchers stick to straightforward and piecemeal theorizing, for example by reshuffling a small number of established concepts, they can be caught in a dilemma. Higher theoretical ambitions could compensate for the lack of resources for elaborate empirical studies, but they may be prevented by the lack of time and knowledge. Furthermore, ‘grand theory’ is not always appreciated in a discipline that favors middle- to low-range theorizing in combination with conventional or elaborate empirical investigations. Works without complex theory, however, can be dismissed as actually ‘atheoretical,’ or ‘not making a theoretical contribution.’ This can be a problem in particular if this impression is not compensated by complex methodologies and large-scale empirical studies.

Still, piecemeal theorizing, often combined with rather heuristic justifications (e.g. by analogy or association), should be reflected critically as it can be particularly prone to false universalism if concepts are taken out of context and generalizations or transfers to other contexts are not carefully reflected. However, such issues can only be really identified by the rigorous reading of specific examples, just as structural issues can only be assessed based on adequate structural data, not by speculative reasoning alone.

In contrast to the aforementioned structural reflections, critical reflection strictly based on the strategies of theory construction as analyzed above can mainly judge whether theorizing is as systematic as possible, in the sense of fully using the potential that different strategies could offer. Although a wider range of modes of reasoning is legitimate in theorizing as opposed to theory criticism or testing, theorizing should ultimately be as structured and rigorous as possible. This does not mean that creativity should be restrained by strict rules from the outset but that the possibilities of theorizing should be ultimately explored as systematically as possible in each case and that theories should be made plausible with the best arguments. As I have argued elsewhere (Krämer, 2022), even if we confront theory with empirical research in one way or another, we should not rely on this confrontation to filter out bad theories because research will be less productive, findings will be less well understood, and practical or critical implications will remain unexplored without the most fitting, fruitful, and as well-justified theories as a starting point.

For example, we should not be contented with ad hoc labels (x can be seen as a type of Y) and ad hoc causal explanation, as used in many of the articles in the sample, that often seem to be based on one-sided intuitions and/or previous research. Rather, free-flowing creativity should pass over into carefully choosing and explicitly applying concepts, a systematic consideration of different causal factors, orders, and mechanisms, and a systematic reflection on the range of theories. To this end, researchers should preferably employ those strategies that ensure a certain completeness and symmetry in the construction of their theories, such as the cross-tabulation of categorizations or a systematic permutation and variation of causal paths. Thus, certain categories and combinations thereof, or causal paths are not cherry-picked or chosen at random. Instead, all counterparts to a category and all possible combinations of categories, or all inverse causal relationships or alternative causal orders will be systematically taken into consideration.

Furthermore, even if space is limited, the theoretical choices should be as well-justified as possible. In particular, the theoretical ‘side-notes’ outside the main theoretical sections – i.e. in introductions, methodological sections, and conclusions – often seem to remain particularly unreflected and unjustified in the articles in the sample. This is where problematic assumptions may be perpetuated and remain unquestioned, or seemingly self-evident ideas may establish themselves: Social media leads to polarization and polarization is a bad thing; reliable and valid causal evidence is only produced by accumulating and reproducing experimental studies; critical scholarship per se goes to the root of things and is therefore more thorough than research in other paradigms, etc.

In the light of a potential theory crisis, good practice in theorizing should be based on further empirical and critical studies and systematic teaching of its methods and would mean to be as systematic and explicit in the construction and justification of theories as possible. The same or similar strategies of systematic variation, recombination, and argumentation should also be used to identify dead angles and biases of previous theories (which also includes a critical reflection of the bias toward cumulativity through theoretical unification that is often inherent in the discussion of scientific crises, see Krämer, 2022).

In many of the articles included in this study, the section on limitations and further research is the least theory-driven or generative of new theoretical ideas – it seems more like a rhetorical device to preempt expectable criticism or a kind of ritual that absolves research from certain shortcomings. Such sections mostly use a number of templates that come down to stating that one has not investigated what one has not investigated: no long-term dynamics in a short-term study, no other countries in a single-country study, not the perspective of other groups when the perspective of one group has been analyzed, etc. In the current case, I have left out and future studies should analyze, for example, other journals and other types of publications such as monographs, actual practices of theorizing from the perspective of researchers and in their daily practice, and theories published in other languages than English.

Two examples may highlight what other strategies of theorizing could be thus identified. If we were to analyze theoretical or theory-heavy monographs (that are often the product of rather privileged positions), we may be able to find ‘architectural’ strategies of constructing more complex overall theoretical frameworks and arguments, as opposed to the often ‘elementary’ strategies identified in the present study dealing with a limited number of concepts and relations (that will, however, be also applied in larger theoretical frameworks). Or if we were to include other outlets further from the dominant pole of the discipline, we could find a wider range of strategies of constructing critical theory, in particular of the variety that goes beyond ‘measuring’ critique in the strict sense (i.e. comparing social reality against some norm), such as disrupting and emancipating critique (Vogelmann, 2017).

However, many of the strategies identified in the sample are usable across types and ranges of strategies. If operations establish, modify, and apply concepts, work with relations between concepts, or imply shifts in intention or perspective are applied, it matters little whether the concepts involved are highly abstract and/or numerous, or part of a few simple theses on very specific phenomena.

However, first, research on theorizing will hopefully continue and study those aspects, and second, theorizing biases and structures that work against certain types of theory will require rigorous theorizing that cannot be done in the final section of an article. Therefore, I will limit myself to the presentation of this study that has been, to my knowledge, of the first of its kind, to calling for more research, and to inviting everyone to critically reflect on our practices of theorizing.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 In the following, ‘theorizing’ is used very broadly, referring to the process of producing theories (Swedberg, 2014a, p. 1), ranging from small changes in existing theories to the construction of complex ones (sometimes, terms referring to production, as in Swedberg’s definition, will be used synonymously for theorizing). Although the terms are rarely defined, different implicit meanings of theorizing and similar terms such as ‘theory building’ or ‘theory construction’ exist, for example with regard to the source of theory production (such as empirical data versus deduction, speculation etc.), the range of theories (general versus limited to specific objects), or the type of theory (e.g., causal, interpretive, or critical). However, the research interest of the present study cuts across these distinctions (on the possibility and potential of an analysis of theorizing across levels of abstraction and types of theories, see also below). Different ways of theorizing in the broad sense will be mentioned subsequently, but they will be designated with more specific terms describing the actual procedures and forms in more detail.

2 References in the results section refer to articles included in the analysis. See Appendix 1 for bibliographical information.

3 Strictly speaking, this ‘should’ can mean different things that are more or less normative or normative in different senses. We may have to do with statements of relevance, even moral obligations or social criticism (e.g., there is a pressing social problem or an oppressed group and it is our duty or morally desirable to conduct research about them and/or let the members of the group speak in their own voice). Or we think we would ‘do’ the object of study ‘justice’ (i.e., produce correct or more adequate statements about it) by researching or modeling it in a certain way. This is only a normative statement in the strict sense if we hold that we are under a kind of moral obligation to hold true beliefs about phenomena or analyze them in the best possible way (not as a means to other ends but as a distinct obligation) (see, e.g., Feldman, 2005).