How to do things with data: Meta-data, meta-media, and meta-communication
First Monday

How to do things with data: Meta-data, meta-media, and meta-communication by Klaus Bruhn Jensen

Moving beyond prevalent technological notions of meta–data, this essay conceptualizes meta–data as the outcome of diverse social practices from the perspective of communication theory. First, I revisit cybernetics and consider the relevance of the concept of meta–communication for the study of the digital media environment. While the bit trails that users leave behind are not normally conceptualized as communication, they are prime candidates for inclusion in a contemporary theory of communication. Second, I discuss digital media as meta–media — media that potentially reproduce and integrate other types of media, old and new. Meta–media yield new varieties of meta–communication and meta–data. Third, I outline a typology of meta–communication for further research on digital media, integrating insights from cybernetics and semiotics.


1. Introduction
2. Two aspects of meta–communication
3. Meta–media and other media
4. Meta–data, connotations, and consequences
5. Conclusion



1. Introduction

The promise of meta–data, including the phenomenon of ‘big data’ (Mayer–Schönberger and Cukier, 2013), for media and communication research is that such data may contribute to ever more comprehensive and granular accounts of media use and communicative practices, across global space and in real time. As such, meta–data complement and, perhaps in certain respects, replace traditional forms of evidence from surveys, content analyses, ethnographic fieldwork, and other empirical approaches. Whereas the field appears unlikely to embrace the notion of ‘the end of theory’ (Anderson, 2008) — the proposition that either interpretive or explanatory theoretical frameworks are made superfluous by the capacity of big–data analyses to map and predict communicative events — media and communication research is currently exploring a variety of new methods to examine the digital media environment (Burgess, et al., 2013). Not only have digital media posed new challenges to research; digital methods, as applied to non-digital forms of communication as well, present new opportunities for empirical inquiry into human communication, culture, and history, as illustrated by developments in digital humanities (Berry, 2012). With more and, potentially, better data at hand, the field of media and communication research might at last, after half a century of scholarship (Park and Pooley, 2008), be in a position to consolidate itself as a discipline, even a science (Berger, et al., 2009). And, by conceiving data as products as well as processes of communication, media and communication studies might provide other disciplines and fields with an interdisciplinary framework for reflecting on and assessing their own practices of collecting, analyzing, and interpreting data.

To begin to assess such promises, I distinguish between two conceptions of data. When processed and understood as vehicles of information and sources of meaning, data can be understood either as representations of reality or as resources for acting in and on reality. The distinction is familiar from classic debates in philosophy regarding the nature of both human cognition and academic research along a continuum from positivism to postmodernism: Do we copy or construct reality through our representations and communications? Feeding into contemporary communication theory, twentieth–century philosophy offered the seminal insight that the everyday use of language — any communication — is a form of construction or action. The title of the present article — ‘how to do things with data’ — recycles the formulation of that insight by John Austin (1962) in his seminal volume entitled How to do things with words. While words, images, and other signs and symbols do provide representations of the world, such representations simultaneously constitute actions, in a context, and for a purpose. A description of the weather may be a simple reminder to bring an umbrella; it may also amount to one turn in a complex debate concerning climate change. Personal information at social network sites can help to identify old friends; the same information can also serve to monitor potential enemies — old, new, and imagined.

The present theoretical essay re–emphasizes the performative aspect, not only of the common practice of communication, but also of the data that communication continuously generates for research. We all do things with words; researchers do additional things with data. In both cases, the actions hold implications far beyond the moment of either communication or analysis. The digital media environment presents new conditions of dialogue and documentation.

The article is divided into three parts. First, I revisit Gregory Bateson’s concept of meta–communication and explore its relevance for the study of online and networked communication. While the bit trails that are accumulated as big data are not normally conceptualized as communication, they are prime candidates for inclusion in a theory of communication that recognizes more of the distinctive affordances of digital media. Second, I note the status of the digital computer, and of computer–based media, as meta–media — media that potentially reproduce and integrate other types of media, old and new. The concept of meta–media sums up the affordances of digital media, simultaneously as means of communication and as conditions of research: Meta–media yield new varieties of meta–communication and meta–data (on various conceptions and origins of the meta–perspective, see Boellstorff, this issue). Third, I rearticulate the mostly technical notion of meta–data in the vocabulary of communication theory, recovering a lineage linking cybernetics and semiotics, and illustrating its significance for future studies of the digital media environment.



2. Two aspects of meta–communication

Bateson (1973) developed the concept of meta–communication with reference to face–to–face interaction, building on diverse studies spanning anthropology and psychiatry. His premise was that, far from being simply the literal exchange of information, “human verbal communication can operate and always does operate at many contrasting levels of abstraction” [1]. Any given statement carries multiple potential meanings that hold implications for the further course of the communication and its outcome. This polysemy is not a contingent product of either channel noise or errors of precision at either end of the channel, but a constitutive feature of human communication. In face–to–face settings, moreover, communication incorporates not just verbal language, but all manner of bodily expressions of the communicators and contingent aspects of their context of interaction. Embodied human beings can be considered media in their own right — media of the first degree (Jensen, 2010). Much of our face–to–face communication is outside our conscious control, and yet, it may qualify as communication as far as other people are concerned. Humans are constantly ascribing meaning to each other as well as to their cultural and natural environments [2]. As students of Bateson summarized the point, humans “cannot not communicate” [3], productively and receptively.

Bateson (1973) identified two aspects of meta–communication. First, people meta–communicate about the codification of their interaction. In some instances, this may amount to meta–linguistic (Jakobson, 1960) or explicit information about the meaning of particular words, phrases, or statements. Taking a standard example from logic — ‘the cat is on the mat’ — Bateson noted that, in a certain context, the reference of ‘cat’ might be said to include tigers, for example, as part of a game. In a wider sense, the codification of communication is in question whenever new topics or vocabularies are introduced into a conversation, and in the common use of humor or irony. In the vast majority of instances, however, people will work out the relevant meaning of statements and expressions as part of the flow of their communication through repetitions, rewordings, examples, etc. Humans are remarkably good at finding out what words and, by implication, other people mean without constantly asking, ‘What do you mean?’

Second, people meta–communicate about their social relationships. In Bateson’s example, the suggestion that a cat is actually a tiger might be an aspect of children’s play, organized role–playing, religious ritual, or other cultural practices. In the unlikely event that someone takes the reference literally and, perhaps, is startled, a meta–communication can reassure them that “This is play” [4]. Again, the nature of the communication mostly becomes clear to people in the course of their turn–taking, including openings, responses, elaborations, etc. It is only rarely that one finds it necessary to explicitly or directly raise questions regarding interpersonal relations or personal identity: ‘Who are you?’, ‘Who am I?’, and, by implication, ‘Who are we?’, as far as this communication is concerned.

In noting the multiple levels of communication, Bateson was further recognizing their interdependence in the practice of communication. For one thing, “the vast majority of both metalinguistic and metacommunicative messages remain implicit” [5] in ‘the message’ as commonly understood. For another thing, “a majority of propositions about codification are also implicit or explicit propositions about relationship and vice versa” [6]. Depending on who one is talking to — family, colleagues, strangers — different kinds and degrees of codification are called for. The meaning of the codes that we use in our communications with each other implies the meaning of our relationship. Both these meanings are established in context.

Bateson (1973) focused on embodied interactions in local contexts. But also in technologically mediated communication across time and space, meta–communication is one constitutive element of the interaction between user and medium, or between user and user. In the case of traditional mass media, codifications and social relationships are signaled and accomplished, not least, by genres: discursive conventions of expressing and experiencing a particular subject matter in common. Whereas the analytical category of genre is most commonly associated with literature, aesthetics, and other humanities, recent decades have witnessed a growing literature on genre simultaneously as a discursive structure and a social practice (Bawarshi, 2000; Lomborg, 2009; Miller, 1984, 1994; Miller and Shepherd, 2004; Yates and Orlikowski, 1992). Beyond epic, dramatic, and lyrical formats, also political negotiations, financial transactions, and cultural criticism constitute discursive forms with social functions. Genres meta–communicate about the anticipated uses of their characteristic contents. For example, news and advertising employ distinctive codes and conventions, and they imply different social relationships with their audiences — as citizens or consumers — who can be expected to interpret the contents as scripts for subsequent action, beyond the moment of communicative interchange.

Digital media join some of the meta–communicative affordances of face–to–face communication and traditional mass communication. On the one hand, digital media, like mass media, rely on a wide variety of genres to address their users — e–mail messages, Web sites, quickpolls, tweets, likes, etc. — but also the genres inherited and remediated (Bolter and Grusin, 1999) from earlier media forms: news, advertising, serial fiction, etc. On the other hand, because of their interactive potentials, digital media reintroduce certain expressive potentials from face–to–face communication — a remediated body language of sorts. Beyond smileys, users enter and leave behind a great variety of information about the communication and about themselves.

Web search engines (Halavais, 2009; Hillis, et al., 2013) offer a useful example of such meta–communication in the digital media environment. Search engines codify information: the algorithms lend a coded structure to an available mass of data points, so that users gain access to them (or not) as information and as this kind of information with some presumed relevance. Search engines also enact communicative relationships: users establish a communicative relationship, not necessarily or explicitly with identifiable individuals or institutions, but with a distributed resource of information, which may lead into multistep flows of communication and interaction, including mundane off–line activities such as finding out where to go to in order to meet someone or buy something. Most important in the present context, in and of their searches, users provide input to the medium or system of communication and, in doing so, they reconfigure the system, however minimally. Such meta–communication prefigures subsequent searches and communications, by the person in question and by others: users participate in the codification of a common information pool, and they articulate identities for themselves vis–à–vis the present information.

It is the continuous documentation of these two aspects of meta–communication through the architectures of digital technologies that, together, present one of the most central challenges to current communication theory. Both information and communicators are codified as part of any act of communication. The question is who codifies what — and whom — with what consequences, when it comes to digital channels of communication and meta–communication. To address that question and challenge, it is useful to return briefly to the definition of digital media. While all media, including humans in the flesh, meta–communicate, digital media are meta–media that meta-communicate in specific ways.



3. Meta–media and other media

The computer can be seen to reproduce and recombine previous media of expression, representation, and interaction on integrated platforms of hardware and software. This unique affordance (Gibson, 1979; Hutchby, 2001) of digital computing was summed up, early on, by Kay and Goldberg (1977), who designated the computer as a ‘meta–medium.’ As modes of expression and representation, digital media recombine text, image, and sound, and they incorporate the full range of traditional genres as inherited from mass media as well as face–to–face interaction: narratives, debates, games, etc. As forms of interaction, digital media integrate one–to–one, one–to–many, and many–to–many forms of communication: in addition to blogging, tweeting, and networking socially, digital media are used extensively for interpersonal contact as well as good old–fashioned mass communication. While the networked personal computer, for a few decades, was the pivotal digital medium in the West, mobile telephones and other portable devices are becoming equally, or more, important access points to the Internet around the world (Castells, et al., 2007). At the same time, the integration of digital technologies into natural objects, cultural artifacts, and social arrangements in a projected Internet of Things (International Telecommunication Union, 2005), once again, is challenging received notions of ‘media’ and ‘communication,’ as illustrated by location–aware technologies and other ubiquitous communication (Silva and Frith, 2012).

Digital communication technologies invite research to focus theorizing not so much on media as artifacts or institutions, but on communicative practices. One key research question is how different flows of communication intersect, on single platforms such as tablet computers, and across several platforms such as smart TVs and cell or mobile phones. Each of these flows entails meta–communication. Their intersections, further, lend themselves to studies of multistep communications, and of the distributed social interactions of which these communications are a constitutive part. To begin, consider some of the prototypical configurations of communication that meta–media enable.


Figure 1: Four prototypes of communication. (Bordewijk and van Kaam, 1986)
Central Allocution Registration
Distributed Consultation Conversation


Figure 1 reproduces the early and helpful typology of telecommunication services by Bordewijk and van Kaam (1986). The conceptual matrix distinguishes, along one dimension, between central and distributed control over an available information base. For example, an encyclopedia or other database may be centrally administered and edited (Encyclopædia Britannica), or it may be open, to varying degrees, to input and revisions from distributed users (Wikipedia). Along the other dimension, it may up to a central authority exactly when the particular items of information become accessible, or it may be up to individual, distributed users. This distinction is illustrated by the difference between broadcast television and various dedicated services, from pay–per–view to streaming services such as Netflix and Hulu.

The two dimensions, in combination, yield four prototypes of communication. Conversation refers to the online equivalents of face–to–face interaction (e.g., chat or conferencing), whereas allocution covers the traditional formats of one–to–many or mass communication, as embedded in digital platforms. Consultation is perhaps typically associated with accessing Web sites and other sources of information, but can also be seen to include, for instance, the more fine–grained process of following Facebook profiles (overlapping, in this systematic, with conversation). Finally, registration concerns the automated documentation of users’ trajectories within a given system of communication, and perhaps beyond, across interrelated media platforms and flows of communication.

It is this last type of interaction — registration, in which users enter information the system, more or less willingly — that speaks most directly to current research and debate concerning meta–data. It is essential to note, however, that registration is not simply equal to the meta–communicative component of digital media. Each of the three other types equally relies on codification and similarly establishes social relationships among users and communication services. The point is that registration represents a distinctive kind of meta–communication with far–reaching implications for other kinds of communication as well as meta–communication on digital platforms. Because codes and relationships are registered in and of various acts of consultation, conversation, and allocution, they lend themselves to later review and analysis by anyone who gains access to the resulting information trail: other users, system administrators, hackers, government officials, etc. In Bateson’s (1973) face–to–face scenarios, communicators were speaking into the air (Peters, 1999). In digital media, communicators are meta–communicating into the system.

One insight of Bordewijk and van Kaam’s (1986) early model was that what had commonly been thought of as separate types of communication, as frequently associated with particular media, were merging or converging on integrated platforms. The model provided a specification of how communication could be said to flow in meta–media. Subsequent technological and institutional developments have made the insights of the model almost common sense to the average media user. My laptop, tablet, and smartphone afford allocution, consultation, conversation, registration, and then some. And yet, the definition and delimitation of meta–media still appear unresolved. In networked media, the flows of communication and meta–communication do not exhibit neat technological or theoretical boundaries. Indeed, the very terminology of meta–media is best thought of as an ad hoc conceptualization of new, digital media with reference to their embedding of old, analog media. In this regard, meta–media and meta–communication are rather different categories of understanding. While meta–media pose a fundamental challenge to a field of research that was founded on a conception of mass media, meta–communication presents a promising avenue for revisiting similarities and differences between old and new media. All media meta–communicate; meta–media meta–communicate in ways that the field is still struggling to account for.

In the next part of the article, I relate Bateson’s (1973) cybernetic concepts to complementary semiotic concepts. Although there are many family resemblances between the broader traditions of cybernetics and semiotics (Heims, 1991; L.E. Kay, 2000; Wilden, 1980), they are not normally articulated in any concrete detail within theories of communication. The development of meta–media represents an opportunity to reconsider the genealogy of the traditions and their current relevance. The following section briefly notes one of the most familiar contributions of semiotics to the wider field of media and communication. I then introduce a somewhat neglected aspect of semiotics, which serves to transpose Bateson’s (1973) general idea of meta–communication onto the concrete analysis of products and processes of communication. The last part of the section revisits the links between semiotics and cybernetics, outlining their combined relevance for the study and use of meta–data.



4. Meta–data, connotations, and consequences

4.1. Connotation languages

If social–scientific research traditions tend to refer to communication in terms of data and information, humanistic traditions rather speak of languages and meaning. The process of translation between these traditions has been ongoing for several decades, and has yielded various forms of integration between transmission models and ritual models, quantitative and qualitative methodologies (for overview, see Jensen, 2012). One particular strength of humanistic traditions has been their sensitivity to the many nuances and multiple levels of meaning that are inherent in human communication, as also highlighted by Bateson (1973). One common limitation of humanistic traditions has been a reluctance to explicate and systematize the analytical procedures employed, and to specify the relationship between the (verbal and visual) ‘languages’ of communication and the second–order ‘languages’ of research. Semiotics stands out as perhaps the most ambitious attempt at joining nuance and precision, and as a prime candidate for developing an interdisciplinary vocabulary regarding the constituents of communication, whether conceived as information or meaning. This has also been recognized in some publications with a social–scientific orientation, for example, in Denis McQuail’s (2010) classic textbook, now in its sixth edition, which includes the application of semiotics to the study of media texts [7].

Roland Barthes’ contribution to semiotics has particular relevance for an understanding of the relationship between communication and meta–communication. However, it is, first and foremost, his conceptualization of denotation and connotation, and of their role in the production of mythologies or ideologies, that has been widely influential in media and communication research so far. Before focusing on meta–communication, then, it is helpful to review the more familiar model and to clarify its complementary relationship with the model addressing meta–communication. In both respects, Barthes (1957; 1973) referred to ‘languages’ — connotation languages and meta–languages — departing from Louis Hjelmslev’s (1943; 1963) formal linguistics.

Connotation languages, first of all, recognize the fact that both individual signs and entire languages are multi–layered structures: one such language can build on another language, giving rise to additional meanings and uses. In modeling connotation languages, Barthes (1957; 1973) sought to specify what exactly is entailed by ‘layers’ of meaning. He went on to suggest that signs and statements accumulate as worldviews — discursive configurations with normative implications. A central motivation of Barthes’ work, and of much semiotics, has been the identification and critique of such worldviews, which could be seen to endorse debatable social conditions as desirable or inevitable.

Figure 2 lays out the principles of a connotation language. The first or bottom level is the level of language as a basic medium of description. The expressive form (signifier) and conceptual content (signified) of a word, image, or statement together constitute a sign that carries a denotation or representational content. The second or top level is the level of the connotation language. Here, the entire sign from the first level is conceived as the expressive form of yet another sign with a conceptual content, sometimes referred to as a connotation. In a further analytical move, a given set of connotations may be interpreted as the constituents of a more of less unified worldview, ideology, or culture. Signs as communicated accumulate as culture.


Language and connotation language
Figure 2: Language and connotation language (adapted from Barthes, 1957; 1973).


To illustrate the principle of connotation languages with one of Barthes’ most famous examples, a magazine cover with an image of a young black man in a uniform saluting the French flag might, in itself, be said to carry a relatively neutral or descriptive denotation. Barthes’ critical point, however, was that the cover immediately elicits a more controversial meaning or connotation. In its historical context, the cover with its content became the expressive form of a further ideological content or myth: ‘French imperialism is not a discriminatory system since people of all colors are in a position to salute the flag, to serve, and to be a part of the nation.’ In sum, this two–tiered mechanism of meaning production could be seen to naturalize certain worldviews, while silencing others.

In later work, Barthes [8] reversed the perspective, suggesting that denotations are not points of departure or foundations of meaning, but rather preliminary end points in a more complex and fluid process of meaning production. Nevertheless, in both perspectives, connotations are considered collective and consequential constructs. Signs will assume a comparatively stable shape when they are used by people for practical purposes of representation and interaction. Contrary to a common misunderstanding, semiotics does not conceive of connotations as the outcome of more or less incidental, personal, or idiosyncratic interpretations. Instead, connotation languages, as employed by individuals and institutions, serve to reproduce and prefer certain worldviews as social facts, for better or worse.

In Bateson’s terminology, connotations might be understood as codifications — the meta–linguistic aspect of meta–communication. Connotations amount to codes that pervade media representations and media users’ frames of interpretation. Simultaneously, codes serve to situate communicators within social relations, a premise shared by Bateson and Barthes. This premise was elaborated by Barthes in a second model paralleling his model of connotation languages. Drawing again on Hjelmslev (1943; 1963), Barthes also referred to meta–languages.

4.2. Meta–languages

Barthes’ use of Hjelmslev’s original terms and concepts was creative, if debatable. Like much of twentieth–century linguistics, Hjelmslev approached languages as systems — systems of communication and second–order systems that either build on or describe such systems. Barthes set out to appropriate — and remediate — Hjelmslev’s basic figure of thought for alternative purposes of analyzing and critiquing contemporary media and communication. Barthes’ accomplishment was to retool the original systemic logic for the study of communication as a practice and a process in social contexts. In addition, Barthes went beyond verbal language to include images and other modalities in his approach to the several aspects and levels of communication.

In Hjelmslev’s (1943; 1963) account, connotation languages and meta–languages have different, but complementary relations to a common reference point, namely, first–order language or language as commonly understood. Connotation languages, on the one hand, add to the meanings of language in such a way that they themselves constitute vehicles of communication. They amount to representations of and statements about the world, as illustrated by Barthes (1957; 1973) with reference to myths such as the notion of an innocent imperialism. Meta–languages, on the other hand, describe language. They are not languages in themselves, but languages about languages, for instance, syntactical or semantic descriptions of the English language. In a wider sense, meta–languages can serve to characterize a language, its uses and implications.

Figure 3 presents the principle of meta–languages. While almost identical to Barthes’ first model of connotation languages (Figure 2), its implications are quite different, underscoring the semiotic insight that small signs, including the details of models, can make a big difference. Compared to Figure 2, Figure 3 inverses the relationship between signifier and signified at the second level. The point is that whereas connotation languages add mostly implicit meanings to the first level of language, meta–languages explicitly address or thematize the first level, adding some further characterization to the basic sign or statement. Its meaning (signified) is left intact, but by adding another description or re–description (signifier), the meta–language invites further consideration of and communication about the first level of language. In Bateson’s (1973) vocabulary, meta–languages may raise questions such as, ‘What is the meaning of X?’, but also, in communicative terms, ‘What do you mean by X?’


Language and meta-language
Figure 3: Language and meta–language (adapted from Barthes, 1957; 1973).


Hjelmslev (1963) had qualified his typology of meta–languages and connotation languages by making a further distinction between scientific and non–scientific languages. Meta–languages could be considered scientific languages. While primarily defined by their formal operations, meta–languages can be operated and applied, above all, by certain expert users, who are able to step back and assume a second–order perspective on the language in question. Such expert users would be linguists, but also, in the case of formal languages and other systems of notation, mathematicians, logicians, or computer scientists. Algorithms amount to languages that depend on meta–languages for their construction and refinement.

At this point, I should clarify how I propose to treat meta–languages, not merely as systems of analysis, but as practices of communication, i.e., as vehicles of meta–communication, which lend themselves to non–expert uses, as well. In doing so, I stand on the shoulders of Barthes, who had performed a creative interpretation of Hjelmslev’s work in order to develop concepts and procedures for the study and critique of communications and their connotations in an age of mass media. In the present age of digital media, meta–languages and meta–communication have taken on an added significance. A new range of meta–languages have become available that afford new forms of explicit and implicit, intentional as well as non–intentional meta–communication. For one thing, the ordinary users of digital media effortlessly employ meta–languages when they tag the blog entries of others; when they customize their profiles at social network sites; and when they forward news stories from Web sites to friends or colleagues. For another thing, the resulting meta–data are accumulated; they can be re–communicated to and by system administrators, advertisers, regulators, and other stakeholders; and the meta–data may inform subsequent actions by any and all of these, with or without the knowledge of the original communicators.

The social uses of meta–data reemphasize Bateson’s (1973) point that meta–communication accomplishes not only the codification of content, but also the maintenance of social relationships. In the digital media environment, these relationships involve users, communication systems, and a host of other social actors who may access and recycle the trails of communication, across space and time, for commercial or surveillance purposes. Far beyond either Batesonian body language or Hjelmslevian language study, meta–communication has major implications for social interactions and structures today. Meta–communication makes a difference.

4.3. How to make a difference

The concept of difference provides a common denominator for the traditions of semiotics and cybernetics. In contrast to the widespread, commonsensical conception of information in both scholarship and everyday parlance as an objective entity or delimited product, the concept of difference entails a relational approach to the constituents and processes of communication. According to Bateson’s [9] definition, “‘information’ may be succinctly defined as any difference which makes a difference in some later event.” From a media user’s point of view, differences manifest themselves, for instance, as messages selected and disseminated (or not) by the press as news; the news of the day as re–communicated (or not) to others face–to–face or online; and actions (or inaction) in response to news of elections, coups, and catastrophes.

Bateson belonged to the tradition of cybernetics — the science of “control and communication in the animal and the machine” (Wiener, 1961). Cybernetics offers formal descriptions of natural as well as cultural processes, which can be seen to share certain elementary structures of information. One further assumption of cybernetics is that these elementary structures enter into highly complex differential structures, from living organisms to information technologies. Bateson drew some of his ideas from one of the two founders of semiotics, Charles Sanders Peirce. Also its second founder, Ferdinand de Saussure, developed a differential and relational conception of information or meaning, and outlined a framework for studying “the life of signs within society” [10].

A fundamental idea of semiotics is that meaning is not a quality of any given sign, but an aspect of how that particular sign differs from other signs. Sounds, letters, and words are defined by the features that distinguish them from other sounds, letters, and words. Sentences, books, libraries, and databases can be examined as increasingly complex configurations of difference. Especially Peirce pursued the idea that the production of difference is not limited to signs or language as such, but carries over to processes of interpretation and social action. In communication theory, one may distinguish between three types of difference [11]:

  • Discursive differences, or the range of potentials meanings of a message: An advertisement is intended to promote some product or service, even if readers may resist or negotiate this meaning (Hall, 1973).
  • Interpretive differences, or the meanings actualized by specific users in particular contexts: The advertisement may be read as consumer information, capitalist propaganda, or everyday entertainment.
  • Performative differences, or predispositions to act: The advertisement may lead readers to either buy or boycott the advertised goods.

The sequential or stepwise aspect of communication, interpretation, and action is of special interest for a reassessment of meta–communication in the context of the digital environment. When registered and documented in digital media, communication lend itself to a great deal of delayed and distributed interpretation and action. One familiar example is ‘liking’ a comment or event on Facebook. This communication entails both of the two aspects of meta–communication: establishing or maintaining a social relationship involving (at least) two communicators, and a codification or valuation of the comment or event in question. In another example, user clicks on a banner advertisement represent interpretations and expressions of interest in a particular (kind of) commodity, leading perhaps to a purchase. Again, a social relationship of buying and selling is enacted, and the representation of the commodity is codified as relevant or attractive from the users’ perspective. In a later stage of meta–communication, the resulting data may be analyzed and resold as evidence of the socio–demographic profiles of potential customers and their codification of relevant commodities. Depending on the kind of commodity, the same data may give rise to further controversial issues. In addition to its constructive uses in agriculture and gardening, fertilizer also has destructive potential as an ingredient of a bomb. A search for or a purchase of fertilizer might variously be codified and socially profiled as a business opportunity or a security threat.

Returning to Bordewijk and van Kaam’s (1986) early typology, one may begin to specify various types of meta–communication. Within their typology, the category of registration covered the documentation of users’ distributed access to different services in a centralized system of communication. In comparison, current interactive systems, while still centralized in key technological and economic respects, afford more degrees of freedom both to users and to system administrators in terms of access to and adjustment of services.

Figure 4 lays out four prototypes of meta–communication, building on Bordewijk and van Kaam’s (1986) model, and reconsidering the differential capacities of systems and users, first, to control the information base and, second, to decide exactly what information will be selected, and when. The prototype most familiar from the original model, akin to registration, is that of processed communication: the documentation and analysis of individual users’ trajectories for purposes of billing, system maintenance, market analysis, strategies of avoiding churn, etc. The prototype diagonally across the model — recommended communication — is comparable, yet different. Here, the focus of attention is on users as collectives or segments, who may or may not continue to communicate as they have done in the past. By reviewing patterns of communication, system administrators can identify favorite acts of communication, as far as different user segments are concerned, but also complements and alternatives that might be recommended to them. While exemplified by recommender services such as those of Amazon for books or Netflix for feature films, the prototype can be extended to include entirely different services from different providers within an extended network of communication. The same logic applies to the marketing and sale of products and services throughout the network to users who are also consumers, and whose communications constitute part of their socio–demographic profile. While the outcome, in Bordewijk and van Kaam’s (1986) terminology, might be thought of as consultation of some communication or commodity, the interaction is different in kind: The offer or recommendation of consultation is the product of an elaborate codification of potential users, who thus enter into a social relationship with the system of communication that is quite different from merely selecting from a database or set menu of news stories, movies, games, or other media ‘content.’


Figure 4: Four prototypes of meta–communication.
System Third–party communication Processed communication
User Recommended communication Iterative communication


Iterative communication is meant to capture the diverse ways in which users interact with reference to each other’s communications. The prototype, again, is similar to Bordewijk and van Kaam’s (1986) category of conversation, but reemphasizes the pervasive iterativity of the conditions of communication in digital media. In addition to opening, participating in, and ending conversations, users comment on, re–send, and act on each others’ communications, whether synchronously or asynchronously and across far–flung networks. In an even wider sense, users are able to shape interfaces and systems as conditions of their communication. Users will customize their points of access to services and networks. They can pull a later push of information to themselves through an RSS feed (Really Simple Syndication). And, they may, to a degree, affect network infrastructures by engaging in collaborative open source innovation (Benkler, 2006; Bruns, 2008; Hippel, 2005). Users, thus, codify themselves and each other as part of variable social relationships.

The last prototype is third–party communication, referring to the accumulation, consolidation, and re–distribution of system–wide evidence of communication patterns. Compared to processed communication, the information base in the case of third–party communication is controlled by the system rather than the user. It is only through the intervention of the system as a technological and institutional agent that a certain kind of consolidated evidence that will be of interest to third parties comes into existence. If processed communication is oriented inward to the operation of the system as such, third–party communication caters outward to additional stakeholders, typically by re–distributing (more or less) refined information to advertisers, marketers, regulators, and government authorities. In Bordewijk and van Kaam’s (1986) typology of communication, allocution or mass communication sends information from a center to a mass of distributed individuals. In the present preliminary typology of meta–communication, the system can be seen to pass on information from (about) individuals to other systems and centers of power.

The motivation for elaborating on these finer details of meta–communication in digital media is that they make a real discursive, interpretive, as well as performative difference. The implications are suggested most prominently by the prototype of third–party communication, but also by iterative communication, both of which speak to ongoing debates about the participatory and empowering potentials of digital media, often along a utopian–dystopian spectrum (e.g., Curran, et al., 2012; Jenkins, 2006). Communication is a form of action in itself; communication also anticipates and negotiates action. In both respects, meta–communication serves to frame and condition communication. By beginning to develop the conceptual resources for assessing the constituents and processes of meta–communication in the present digital media environment, research may itself make a small difference to future communication practice and communication policy.



5. Conclusion

This theoretical essay has sought to contribute to a better understanding of meta–data as constituents and processes of digital information systems from the perspective of communication theory. Meta–data can be conceived as particular varieties of meta–languages (Hjelmslev, 1963), which have been enabled by the development of meta–media (Kay and Goldberg, 1977) that hold a unique set of combinatorial and interactive affordances. Relying on and redeveloping these affordances, both media users and systems of communication are in a position to engage in a new range of meta–communicative practices (Bateson, 1973), from ‘liking’ and recommending communication among users, to systems of personalized advertising and government surveillance.

To restate a key point, meta–communication is not a special feature of digital media, but a characteristic of all human communication in any medium. Digital media, however, have introduced new capacities for meta–communication with significant structural implications. Having examined interpersonal communication and organizational communication with some reference to meta–communication, and having explored genres as implicit meta–messages of mass communication, it is high time that the field of media and communication studies turn its attention to meta–communication as a central area of theoretical and empirical research concerning digital media. In this regard, the field also has a further contribution to make in other disciplines and fields currently in need of a framework for reflecting on and assessing practices of collecting, analyzing, and interpreting (big) data. The present article has only been able to outline a preliminary typology, and to consider relevant conceptual family resemblances across cybernetics and semiotics. Together, the two traditions may serve to theorize meta–data and meta–communication as social phenomena and practices, above and beyond their most prevalent conceptions in terms of technological infrastructures and economic transactions.

Beyond a critical review of the proposed typology itself and an operationalization of its elements for empirical inquiry, further research is needed on the relationship between meta–communication and communication as traditionally understood, whether as a transmission or as a ritual (Carey, 1989), when it comes to digital media. As emphasized by Bateson (1973), the multiple levels of language and communication interact. Codifications, but also contents, as sent or received, bear witness to the orientations, norms, and values of the communicators. In the terminology of Hjelmslev and Barthes, meta–languages intersect with connotation languages. We are defined socially by who we communicate with, in what codes, but certainly also by the content of our communications and their place in wider networks of meaning and community. It is for this reason, not least, that the trails of meta–communication motivate big–data analyses by businesses and security agencies alike. It is also for this reason that so many users spend so much time and money meta–communicating among themselves. We all do things with data in the digital media environment. End of article


About the author

Klaus Bruhn Jensen is Professor in the Department of Media, Cognition, and Communication, University of Copenhagen, Denmark, and Vice Head of its Center for Communication and Computing. Life Member for Service of the Association of Internet Researchers and a Fellow of the International Communicology Institute. Recent publications include contributions to the International encyclopedia of communication (12 volumes, Blackwell, 2008 —, for which he serves as Area Editor of Communication Theory and Philosophy; Media convergence: The three degrees of network, mass, and interpersonal communication (Routledge, 2010); and A handbook of media and communication research: Qualitative and quantitative methodologies (second edition, Routledge, 2012).
E–mail: kbj [at] hum [dot] ku [dot] dk



This article builds on and extends arguments first published in chapter 5 of Jensen (2010).



1. Bateson, 1973, p. 150.

2. Ruesch and Bateson, 1951, p. 6.

3. Watzlawick, et al., 1967, p. 49.

4. Bateson, 1973, p. 151.

5. Ibid.

6. Ruesch and Bateson, 1951, p. 209.

7. McQuail, 2010, pp. 345–350.

8. Barthes, 1970, pp. 13–16.

9. Bateson, 1973, p. 351.

10. Saussure, 1959, p. 16.

11. See Jensen, 2010, pp. 44–47.



Chris Anderson, 2008. “The end of theory: The data deluge makes the scientific method obsolete,” Wired, volume 16, number 7, at, accessed 15 May 2009.

John L. Austin, 1962. How to do things with words. Oxford: Clarendon Press.

Roland Barthes, 1973. Mythologies. Selected and translated by Annette Lavers. London: Paladin Books.

Roland Barthes, 1970. S/Z. Paris: Éditions du Seuil.

Roland Barthes, 1957. Mythologies. Paris: Éditions du Seuil.

Gregory Bateson, 1973. Steps to an ecology of mind: Collected essays in anthropology, psychiatry, evolution and epistemology. London: Granada.

Anis Bawarshi, 2000. “The genre function,” College English, volume 62, number 3, pp. 335-360.

Yochai Benkler, 2006. The wealth of networks: How social production transforms markets and freedom. New Haven, Conn.: Yale University Press.

Charles R. Berger, Michael E. Rolof and David R Roskos–Ewolen (editors), 2009. The handbook of communication science. Second edition. Newbury Park, Calif.: Sage.

David M. Berry (editor), 2012. Understanding digital humanities. New York: Palgrave Macmillan.

J. David Bolter and Richard Grusin, 1999. Remediation: Understanding new media. Cambridge, Mass.: MIT Press.

Jan L. Bordewijk and Ben van Kaam, 1986. “Towards a new classification of tele–information services,” InterMedia, volume 14, number 1, pp. 16–21.

Axel Bruns, 2008. Blogs, Wikipedia, Second Life, and beyond: From production to produsage. New York: Peter Lang.

Jean Burgess, Axel Bruns, and Larissa Hjorth (editors), 2013. “Special issue: Emerging methods for digital media research,” Journal of Broadcasting & Electronic Media, volume 57, number 1, pp. 1–114.
doi:, accessed 20 September 2013.

James W. Carey, 1989. “A cultural approach to communication,” In: James W. Carey. Communication as culture: Essays on media and society. Boston: Unwin Hyman, pp. 13–36.

Manuel Castells, Mireia Fernández–Ardèvol, Jack Linchuan Qiu and Araba Sey, 2007. Mobile communication and society: A global perspective: A project of the Annenberg Research Network on international communication. Cambridge, Mass.: MIT Press.

James Curran, Natalie Fenton, and Des Freedman, 2012. Misunderstanding the Internet. London, New York: Routledge.

James J. Gibson, 1979. The ecological approach to visual perception. Boston: Houghton Mifflin.

Alexander Halavais, 2009. Search engine society. Cambridge: Polity.

Stuart Hall, 1973. Encoding and decoding in the television discourse. Stencilled occasional paper, SP number 7. Birmingham, U.K.: Centre for Cultural Studies, University of Birmingham.

Steve J. Heims, 1991. The cybernetics group. Cambridge, Mass.: MIT Press.

Ken Hillis, Michael Petit, and Kylie Jarrett, 2013. Google and the culture of search. New York: Routledge.

Eric von Hippel, 2005. Democratizing innovation. Cambridge, Mass.: MIT Press.

Louis Hjelmslev, 1963. Prolegomena to a theory of language. Revised edition. Translated by Francis J. Whitfield. Madison: University of Wisconsin Press.

Louis Hjelmslev, 1943. Omkring sprogteoriens grundlæggelse. København: B. Lunos bogtrykkeri a/s.

Ian Hutchby, 2001. Conversation and technology: From the telephone to the Internet. Cambridge: Polity.

International Telecommunication Union (ITU), 2005. “The Internet of things: Executive summary,” at, accessed 28 March 2008.

Roman Jakobson, 1960. “Closing statement: Linguistics and poetics,” In: Thomas A. Sebeok (editor). Style in language. Cambridge, Mass.: MIT Press, pp. 350–377.

Henry Jenkins, 2006. Convergence culture: Where old and new media collide. New York: New York University Press.

Klaus Bruhn Jensen (editor), 2012. A handbook of media and communication research: Qualitative and quantitative methodologies. Second edition. New York: Routledge.

Klaus Bruhn Jensen, 2010. Media convergence: The three degrees of network, mass, and interpersonal communication. London: Routledge.

Alan Kay and Adele Goldberg, 1977. “Personal dynamic media,” IEEE Computer, volume 10, number 3, pp. 254–263; reprinted in: Paul A. Mayer (editor), 1999. Computer media and communication: A reader. Oxford: Oxford University Press, pp. 111–119.

Lily E. Kay, 2000. Who wrote the book of life? A history of the genetic code. Stanford, Calif.: Stanford University Press.

Stine Lomborg, 2009. “Navigating the blogosphere: Towards a genre-based typology of Weblogs,” First Monday, volume 14, number 5, at, accessed 1 September 2013.

Viktor Mayer–Schönberger and Kenneth Cukier, 2013. Big data: A revolution that will transform how we live, work, and think. Boston: Houghton Mifflin Harcourt.

Denis McQuail, 2010. McQuail’s mass communication theory. Sixth edition. London: Sage.

Carolyn R. Miller, 1994. “Rhetorical community: The cultural basis of genre,” In: Aviva Freedman and Peter Medway (editors). Genre and the new rhetoric. London: Taylor & Francis, pp. 67–78.

Carolyn R. Miller, 1984. “Genre as social action,” Quarterly Journal of Speech, volume 70, number 2, pp. 151–167.

Carolyn R. Miller and Dawn Shepherd, 2004. “Blogging as social action: A genre analysis of the Weblog,” In: Laura J. Gurak, Smiljana Antonijevic, Laurie Johnson, Clancy Ratliff, and Jessica Reyman (editors). Into the blogosphere: Rhetoric, community, and culture of Weblogs. Minneapolis: University of Minnesota, at, accessed 1 September 2013.

David W. Park and Jefferson Pooley (editors), 2008. The history of media and communication research: Contested memories. New York: Peter Lang.

John Durham Peters, 1999. Speaking into the air: A history of the idea of communication. Chicago: University of Chicago Press.

Jurgen Ruesch and Gregory Bateson, 1951. Communication: The social matrix of psychiatry. New York: Norton.

Ferdinand de Saussure, 1959. Course in general linguistics. Translated by Wade Baskin. London: Peter Owen.

Adriana de Sousa e Silva and Jordan Frith, 2012. Mobile interfaces in public spaces: Locational privacy, control, and urban sociability. New York: Routledge.

Paul Watzlawick, Janet Helmick Beavin, and Don D. Jackson, 1967. Pragmatics of human communication: A study of interactional patterns, pathologies, and paradoxes. New York: Norton.

Norbert Wiener, 1961. Cybernetics, or control and communication in the animal and the machine. Second edition. Cambridge, Mass.: MIT Press.

Anthony Wilden, 1980. System and structure: Essays in communication and exchange. Second edition. London: Tavistock.

Joanne Yates and Wanda J. Orlikowski, 1992. “Genres of organizational communication: A structurational approach to studying communication and media,” Academy of Management Review, volume 17, number 2, pp. 299–326.


Editorial history

Received 16 September 2013; accepted 17 September 2013.

Commons licens
“How to do things with data: Meta–data, meta–media, and meta–communication” by Klaus Bruhn Jensen is licensed under a Creative Commons Navngivelse–IkkeKommerciel–IngenBearbejdelse 3.0 Unported License.

How to do things with data: Meta–data, meta–media, and meta–communication
by Klaus Bruhn Jensen.
First Monday, Volume 18, Number 10 - 7 October 2013

A Great Cities Initiative of the University of Illinois at Chicago University Library.

© First Monday, 1995-2014.