Codes Colliding in Connective Cultures

The Emergence of the Norwegian Police Emergency Control Room Twitter

1_Code in Connective Cultures [1]

Codes manifest in connective cultures in complex and often contradictory ways. Codes simultaneously shape and are shaped by how humans communicate and navigate across digital and analogue landscapes; they connect across borders and create boundaries within groups. At the same time, code materializes these very landscapes. Computationally speaking, code constitutes the digital spaces where we communicate through software: as languages of programming, code takes active part in communication in connective cultures. The digital transformation of society has implications both for public institutions and the private lives of individuals. Significantly, the increasing impact of social platforms has brought about a shift in the relations between the public and the private, through creating sites where codes of communication and conduct previously existing in separate spheres meet, or, more often, collide. [2] The presence of police services on social media platforms is one such site. In this article, I will use the concept of code as a tool to make sense of phenomena arising from digital spaces where, as a word, code traditionally has different meanings. Exploring the Norwegian Police Service’s use of Twitter as a cultural phenomenon, I ask: how can we utilize the inherent multiplicity of the word code as a conceptual tool to bring out the complexities of cultural phenomena in connective cultures?

As part of the digital, code is commonly understood quite concretely as a tool for programming, while in the analysis of the cultural, code is more often understood semiotically as part of a sign system, describing the learned conventions that govern processes through which humans make sense of their worlds. It can also be understood as codes of conduct, formal or implicit rules governing practices within groups. The difference is captured by Yuri Lotman in his description of the relative nature of semiotic systems, as he asserts “[…] a code includes not only a certain binary set of rules for encoding and decoding a message, but also a multi-dimensional hierarchy.” [3] In the technical sphere, code is indeed most often used as an expression of the binary set of rules for encoding and decoding that characterize the language used for programming. Yet in connective cultures, these codes become intertwined with natural language as the digital and cultural worlds integrate, and as such become part of the multi-dimensional hierarchy of meaning-making.

I use the term connective cultures as a description of cultures where digital means of communication have become integrated into everyday life in such a way that it is meaningless to define an inside or an outside of digital communication technologies. I draw on José van Dijk’s concept of a culture of connectivity, viewed from a different perspective. [4] While culture of connectivity highlights the impact of platforms and the businesses that run them, i.e., the culture that emerges from connectivity, connective culture shifts focus to cultures where connectivity is part of everyday life, and where their mutual shaping of each other is entangled and indivisible. This article explores what emerges when the different meanings of the word ‘code’ meet through the Norwegian Police Service’s presence on Twitter. I read this presence as a cultural phenomenon: that is, not as a social phenomenon emerging from the structures of society, but as a site where meaning-making takes place. Focusing on the presence of a state organization on Twitter rather than just Twitter as a platform or a general topic allows for exemplifying more clearly the entanglements that arise in contemporary connective cultures where the codes governing public and private merge in new ways.

The use of code as an analytical concept in this article addresses the new entanglements that arise from lives lived with digital platforms. While I draw on cultural semiotic understandings of code as a specific function of language, [5] I propose expanding the understanding of the concept using insights from a material-discursive framework following Karen Barad’s notion that matter and meaning are entangled, not as separate entities that coexist and interact, but as a mutually constitutive agency in intra-action. [6] The neologism intra-action is an alternative to interaction: phenomena do not consist of previously existing stable entities that act on one another, but of agencies that should be understood as emerging from phenomena. Meaning and matter are not determined through some pre-existing state, but emerge from phenomena as they unfold. [7] In the context of this article, this entails that we cannot analyze the meanings emerging from connective cultures as pre-conceived, stable codes that humans use to make sense of their environments. Rather, the meanings that emerge from connective cultures are effects of humans intra-acting with the technologies that materialize communication and the wider material-discursive aspects of their cultures. In this sense, codes are not only the tools that humans use, whether computationally to manipulate data and software or culturally to make sense of their lives, they are also agents that co-constitute and are themselves constituted by the meaning-making process. Code is language and materialization all at once.

2_Police and Social Media

The Norwegian Police Service entered Twitter in 2011, when the Oslo police district registered its official account. Several other police districts followed suit, and today, all twelve police districts in Norway have their own emergency control room-based Twitter accounts where they report (mainly) selected events and (sometimes) general concerns. The Norwegian police use several different social media platforms in their work, and they have different strategies for the Twitter accounts run by the emergency control rooms versus the Twitter accounts run by police stations, or for accounts on other platforms such as Facebook, Instagram, SnapChat, and TikTok. The emergency control rooms regard Twitter as a strictly informational channel, the purpose of which is to inform media and the public about ongoing events, and police-educated staff write all posts. This means that the tweets most often are short and descriptive, without the use of multimedia or emojis. The other platforms are run either by operational or other staff at police stations, or by online police patrols: police-educated staff who are specifically hired and trained for a role on online platforms that is similar to that of street patrols, with the intent of promoting dialogue with the public. [8] A crucial point in this regard is that programmatic code is not one thing: social media is not one entity, but a coded sphere where different platforms have different contingencies, creating different entanglements with specific effects. The Norwegian police on Twitter is not the same as the Norwegian police on Facebook, or Instagram, or TikTok. What makes the emergency control room Twitter a particularly interesting case in this context is that it is not used as a social network, but as a tool for disseminating information. This makes it a site where the inherent contradictions between codes governing the institutional and the social become visible, as will be discussed below.

Previous research on police and social media in an international context largely focuses on asking if the police use social media effectively to fulfill their role in policing. A case in point is the 2011 article by Jeremy Crump titled “What are the police doing on Twitter?,” which concludes that the police should be doing something else. [9] Reading this through the concept of codes, the majority of research takes institutional code as its vantage point (what should the police be doing?), and then investigates to what extent police use social media platforms (the affordances of the coded environments, the cultural codes regulating behavior and language use) in an effort to help them fulfill their role. However, this practical focus cannot hide the inherent contradictions of the presence of police on social media. While the focus is on whether the police follow their institutional code, a general premise is that social media are spaces for public engagement and dialogue. The police are supposed to be what they are, while also doing what one should do on social media. These demands are coded differently in society, but they exist simultaneously on social media platforms. Christopher O’Connor and Maggie Dwyer, writing about Canada and Kenya respectively, point out that police use Twitter to connect with and engage the public. [10] Stephan G. Grimmelikhuijsen and Albert J. Meijer are critical towards the Dutch police’s use of Twitter instead of other platforms, as they find that Twitter is a platform that does not encourage dialogue. Karen Bullock and Jeremy Crump, writing a few years apart, find the same in the UK. [11] Thus, while there is a complex intertwinement of codes at play in the phenomenon of police on social media, the conclusions in present studies most often tend to discuss this presence according to the ways that police adjust to the codes of social media. In this article, I suggest that shifting focus from what police are doing, to how their presence plays out as an ongoing negotiation of codes, can bring deeper insight into the phenomena of public institutions on social platforms. [12]

Across social-science research studying police and social media, there are already in place relatively stable conceptions regarding what ‘the police’ is, what ‘social media’ is, and what the police should be doing on social media. These can be conceptualized as codes, determining how researchers interpret the phenomena and the conclusions that emerge. However, I argue that these codes should not be applied as stable reference points, but as flexible parts of emerging phenomena that take on different meanings in their intra-action with other phenomena. In the following, I will use the concept of codes to explore the entanglement of the Norwegian police service on Twitter, to create both a foundation for understanding the Norwegian police in connective culture, and a foundation for reading this phenomenon through other similar phenomena that emerge in connective cultures.

3_Close and Distant Reading

The exploration is based on a dataset consisting of all tweets made by Norwegian emergency rooms between September 28, 2011 and August 14, 2021 (N=453.610), as well as official documents concerning the use of social media by the Norwegian police, and selected articles from news media. In other words, my main source is text, read both as materiality (using programmatic code), and as meaning (drawing on code as a cultural semiotic concept), and analyzed within a framework where these dimensions might have alternating weight but can never be separated. This methodology has touching points with research that aims to rethink the role of text in history, such as Kristin Asdal and Helge Jordheim’s theorization that reads Ferdinand Saussure’s structuralism through the insights of Actor-Network Theory (ANT) as developed by Bruno Latour. [13] Asdal and Jordheim argue that texts are mobile, they are not bound by a certain context, but move through history and time, and they move their recipients. Although my analytical focus in this article is on the concept of code, the mobility of the texts in question is inherent to the analysis. I aim not to establish the correct code through which police-Twitter emerges, but to establish that codes are emergent, which implies that they are bound by the continuously changing contingencies of text and practice: not inherent, not arbitrary, but contingent and always on the move.

The dataset was prepared for computational analysis by removing all signs that were not letters, and making all letters lowercase. The dataset contained no images or videos and very few emojis, which likely expressed the intent of the tweets to be sober information. The words were then lemmatized, which means that all words were changed to their basic grammatical form, i.e. cars are turned into car, better and best is turned into good. [14] Finally, I removed stop words, i.e. words that occur frequently, but that are not considered important for the creation of topics, such as and, in, and also; as well as context-specific words that are used so often they have no relevance for forming topics, such as the police phone number for tips and general inquiries (02800). Text preparation alters the material in a specific way, based on a significantly human idea about what the codes should reveal. Already here, the semiotic and material is intertwined. The short and dry descriptions of the procedures followed in preparing the dataset for analysis undermines the work that goes into this process: the co-creation of a workable dataset is a time-consuming labor, demanding both the human and computational brains, that in the end creates the material foundation for the analysis.

The tweets were first analyzed computationally using topic modeling. Topic modeling is a Natural Language Processing (NLP) method that indicates topics in a text or corpus of texts by creating lists of words that are likely to appear in relation to one another. [15] A human must then interpret these lists to decide whether they are meaningful in a human context. There are several different algorithms available that will yield different results. I experimented with different types, before settling for GSDMM, an algorithm considered particularly suitable for tweets because it assumes that each document (each tweet) consists of a single topic, generating output that better conforms to the current human expectations of a tweet, and thus produces more ‘meaningful’ results, given the context. [16]

Running the same algorithm with the same parameters on the same corpus will not yield the exact same result every time. The most stable part of the entanglement is perhaps the human, who does the decoding of the topics based on their preconceptions of what these topics should be. Indeed, when running my experiments, I produced patterns of topics early on across different approaches. The algorithm produced different lists, yet my context of interpretation produced a limited set of topics; the material code and data produce a list of words, and the semiotic code and the human decide what to include and exclude, then naming the inclusions. All of this—not only the relation between algorithm and tweets, but the entire material-discursive context that has conditioned the codes through which the human researcher makes their interpretation—is part of the entanglement that creates the phenomenon. This method is thus a pertinent example of how meaning is created through material-discursive relations within the experiment, and not an inherent property of the text that is lifted from it through mathematical calculations and named by an all-knowing human. Computational analysis creates patterns, but we should understand these as figurations for helping us think about a certain phenomenon, and not as representations of an externally existing phenomenon.

The results of this topic modeling provided the basis for my choosing tweets for close reading. I did not read all 453,610 tweets, but chose segments based on dates and periods, and by searching for words and topics based on the results of the topic modeling. Of course, the material tweets I worked with were different from the material tweets the algorithm worked with—I read the tweets as they appeared when harvested from the Twitter API. These were read in succession and summarized based on my contextual and cultural vantage point. The idea was to create a synthesis of how I decoded these tweets in successive reading, highlighting differences, not similarities. In addition to this, I collected texts from news media that were about the police’s use of Twitter and the Police Directorate’s official guidelines for language and content on Twitter.

In the following, I will first address the possibilities of using code as a concept to understand aspects of the police on Twitter and discuss what emerges when programmatic codes are set to read the text as data. What is the phenomenon of the Norwegian police on Twitter as I read the measurements made by the apparatus of a GSDMM algorithm? Then I continue on the level of singular tweets, the text-as-language, if you will. Here, I use code interchangeably as a tool to identify emergences from the entanglement, and as the object under study through exploring the role of code in producing these emergences, elucidating that cultural phenomena are entanglements far more complex than just a combination of preceding entities. Understanding the police on Twitter is not simply a question of understanding how they navigate codes, but understanding that the practice in itself is a separate phenomenon with its own effects, which from here forward I will describe as police-Twitter.

4_Code and Topics

Tweets are meaning and matter. They are figurations of language containing semiotic meaning, but they are also material data, and in their entanglement, they emerge in myriad ways in the everyday lives of humans all over the world, and they are productive in just as many ways. Meaning is encoded into their expressions as language in use, but it is also encoded in how and where they appear as coded data in a network, algorithmically shared with followers, found through manual searches, liked, responded to, re-tweeted, appearing on computers, smart phones, iPads, on the bus, in the bed, at work in the emergency control room. In the following, I will explore code as it works and emerges through computational analysis of tweets. How can we understand these computational codes as intertwined with semiotic code within cultural phenomena in general, and police-Twitter in particular?

The coded environment of social media provides a material ground for analyzing overall discourses. Simon Lindgren argues that the datafication of society means that humanities and social sciences should strive to move beyond computational methods, towards new and creative theorizations. [17] Over the last decade computational methods, under the rubric of Big Data and Artificial Intelligence (AI) has been hailed by some as the solution to everything from crime prevention to city planning, while at the same time being demonized and feared by others. This hope/fear-spectrum is tied to an assumption that given the size of the data, the methods in and of themselves will necessarily produce fact. [18] In a seminal text, Kate Campbell and danah boyd effectively picked apart the dream about all the things that big data can provide, concluding that “We should consider how the tools participate in shaping the world with us as we use them.” [19] Similarly, Lindgren argues “[…] data need theory, for considering both the data, the methods, the ethics and the result of the research.” [20] From a material-discursive point of view, this means that computational methods do not produce objectively truthful accounts, only phenomena with highly specific effects that are always already entangled in theories and discourses. What results from computational analysis is always already an intra-action involving human concepts.

Not only are social media places where semiotic codes abound, they are also places where communications materialize as data that we can download and analyze as material code. Furthermore, the communication occurring is already constrained by the coded environments of that platform’s affordances. [21] Each platform is built with its own possibilities and restraints, whose production is an entanglement of technological possibilities and negotiations over what these platforms are and should be. [22] In the becoming of social media, semiotic code shapes programmatic code shapes semiotic code shapes programmatic code, in a continuous and simultaneous process that looks linear only when written on the page. Topic modeling provides a look at this entanglement by way of topics that comprise a pre-set number of words for the researcher to decode. The programmatic code of the algorithm performing the modeling is entangled with the semiotic codes involved in the researcher’s interpretation of the data. Thus, the emerging topics do not pre-exist the experiment, they are created as the algorithm, data, and researcher mutually intra-act, becoming entities that highlight aspects of how the entanglement of police-Twitter is encoded.

5_Traffic, Fires and Disorderly Conduct

The overall topic that emerged from my decoding of the topic modeling of police-Twitter is “mundane events.” Repeatedly, I decoded topics as traffic, fires, and miscellaneous disorderly conduct in the public sphere and at private addresses. There were instances of dogs lost in the inland districts, and often boat-related incidents along the coast. In 1974, Egon Bittner defined police work as ‘something-that-ought-not-to-be-happening-and-about-which-someone-had-better-do-something-now!’ [23] Of course, ‘police’ is in itself an entangled phenomenon, and is different phenomena in different locations and contexts, yet this fairly old definition is still descriptive of the police as they emerge through Twitter. The phrase mundane events is something of an oxymoron; is it really an event if it is mundane? Yet there is already something within the semiotic landscape of policing that underlines that, in this sphere, the mundane is always already an event. The police exist because things regularly happen in society about which someone had better do something. The materialization of events through tweets from the emergency control room suggests that it is not the extraordinariness, but the commonness of such events that condition this form of police work. Police work emerges as events connected to everyday life, not high-profile crimes: Fires are extraordinary for those who experience them, but they do happen regularly, and they are to be expected. Traffic incidents are extremely common, they most often create congested roads, material damage and stories: the police traditionally show up to deal with the first two of these; in the age of Twitter, the last aspect is also integrated.

Read as a summary of main topics, topic modeling could be seen as providing a materialization of the overall and likely implicit impression of what the police do, obtained by  people who follow the police Twitter account. This is what becomes the issue for one who follows the police on Twitter, algorithmically receiving their updates or being actively aware of the police Twitter account as a place to go to find out what is happening. [24] At the same time, police-Twitter emerges as a phenomenon though these practices. Police-Twitter becomes encoded as a place to get answers when something is happening in one’s vicinity, which means that police on Twitter emerge as a phenomenon with a specific meaning through their practices of tweeting. The level of meaning and the level of practice is mutually constitutive in the becoming of the phenomenon; code is meaning-making in practice.

The issues that emerge look like those identified by Jenny Maria Lundgaard in an ethnographic study of the Norwegian police emergency control room. [25] Lundgaard elucidates that operators in the emergency control room play an important part in establishing what is defined as police work, and what is dismissed as noise. Although invisible to the recipients of the tweets, the entanglement that is the emergency control room is an important agent in the becoming of the tweets, though producing different effects. In the control room setting, the intra-action is a messy affair that produces police work: it creates events and makes things happen. On Twitter, the emergency control room becomes a different agency in intra-action with a coded landscape distributed in the population. Still intimately tied to the agencies that create practical police work, the control room and all its messy workings are excluded from the phenomenon as it is decoded by the population. As the agency of the control room becomes entangled with the coded landscape of Twitter, a different agency emerges, coded as the types of events to which police respond, to protect and help us: traffic, fires, and the myriad of ordinary out-of-the-ordinary incidents that might happen in everyday lives.

6_Code and Texts

What emerges from topic-modeling the emergency control rooms on Twitter is a familiar description of existing knowledge about events that materialize through the emergency control room. [26] Yet it is not a given that the codes of Twitter and policing should coexist in such a harmonious way. Reading the insights from text-as-data through the tweets as texts, it appears that police-Twitter is not a straightforward case of engaging or informing the public. Rather, it is an entanglement that produces effects independent of and beyond the intentions of the police, the affordances of the social media channel, and the corpus of tweets on the platform. In the following I will present insights from a close reading of the police tweets, and will read these through the insights emerging from the preceding computational analysis.

Twitter is not an official police channel, it is a media company and a social sphere, a phenomenon with its own dynamic codes. Twitter is a vastly entangled phenomenon, but in the contemporary Norwegian context, Twitter can be described as a forum for debate dominated by middle-aged academics, active debaters, the media, and people from the so-called “cultural elite” (a description most often made by others). [27] Their expression is characterized by strong opinions and use of humor and satire. However, the coded affordances of the platform assure that possibilities of emergences and new entanglements are open. The constraints for actual access to Twitter are thoroughly material-discursive: you need at least some form of computer, an Internet connection, some basic knowledge of how to navigate a browser, and to be in a country where access to Twitter is not compromised by the authorities (or, alternatively, the technical skills to enable circumvention of such a ban). What Twitter is at any given moment is constantly emerging, and police-Twitter is one of these emergences, through, to quote Judith Butler, “a process of materialization that stabilizes over time to produce the effect of boundary, fixity and surface.” [28]

I read the police’s first fumbling keyboard taps on Twitter as attempts to find an identity in this landscape, navigating the cultural codes of social media communication and the institutional codes of policing. In their first few months on Twitter, this sometimes resulted in tweets that stand out when viewed with today’s lenses: several of the first tweets to come out of Oslo, for instance, were signed with the name and title of the control room sergeant, and the fifth tweet ended with a smiley: “the traffic incident has been cleared, traffic is flowing nicely again :-).” [29] The use of name and title in tweets could be an attempt to express police authority, or it could be a pragmatic decision tied to the fact that the police entered Twitter to inform news media about events more efficiently, thus handing them a quote for their news report. [30] Whatever the reason, when decoded as a tweet, the practice is uncommon, and arguably connotes an unskilled user. To the general reader, the names seem to appear in random tweets, following no apparent logic, and after a few weeks they stopped appearing altogether. The smiley is more of an effort to adjust to the general codes of the Twittersphere, yet it sits oddly within an informational tweet about traffic. The two examples highlight that there are two sets of meaning-making practices entangled in the police presence on twitter, two sets of codes that flow together. However, these easily noticeable examples of grappling with codes were in the minority from the start. Close reading supports the impression from topic modeling that, early on, the Norwegian police used Twitter as an informational channel for events, with tweets like “Fire in car close to Ammerudhellinga, extinguished, but likely intentional. No casualties. Searching the area for suspects,” and “E6 northbound, before Karihaugen. Chain collision. Police are underway.” [31] However, in looking for these subtle differences that break with expectation, it becomes clear that the system of codes is not connected to either the codes of the police or the codes of Twitter, but that they comprise a new set of codes. The phenomenon is not understood through simply distinguishing preceding codes but highlighting that what emerges is a separate set of codes that flow together, producing a different phenomenon.

7_Humor and Danger

As a cultural phenomenon, police-Twitter is entangled with aspects of culture beyond the codes of the police and the tweets as they appear on the coded landscape of the platform. Traditional media is another productive part of this entanglement. On February 5, 2013, the local Oslo version of the national newspaper Aftenposten reported that the Oslo police had won the award for ‘funniest tweet of the year’ at the national conference Social Media Days. [32] The tweet read “Storo: we had a call regarding ongoing noise and women screaming. When we arrived, we found nurses having a warm-up party. We will leave the premises soon.” [33] The article focuses on Oslo police district’s use of humorous tweets, and the attention that they have gained for this. As the tweet shows, this is not an outright joke. “We will leave the premises soon” is a play on the more common ending stating that they have left the premises, insinuating that they are in no hurry to leave a fun party. The explicit mention that these are nurses having a party can also be read as a playful reference to the prejudice that police are men and nurses are women. This tweet exemplifies a sophisticated use of code, combining the informative police code with the playful double entendre of Twitter. Although one control room sergeant is quoted in the newspaper article saying that they usually use Twitter as a serious channel for information, this point is drowned out in praise for their informal tone.

The following day, the same newspaper turned the tables with the headline “It is dangerous to be funny on Twitter.” [34] This article asks if the use of humor might damage the police’s reputation and addresses the potential for unwittingly disclosing details that are in breach of the involved parties’ privacy. This risk is also explored by Andrew Goldsmith in a study of the Australian police’s use of social media, wherein he considers several cases where the affordances of social media might unwittingly lead to indiscretion. [35] The issue of the police being funny on Twitter emerges from time to time in Norwegian media. What is interesting in our context is not necessarily the argument that this might damage the police’s  reputation, but the fact that the Norwegian police’s actual use of humor on Twitter, empirically speaking, is not very common. The idea that the Norwegian police are funny on Twitter, however, is widespread. Use of humor is not part of the institutional code of police conduct, but emerges from its entanglement with codes attached to Twitter. The occasional use of humor stands out, amplifying these tweets through creating qualitative effects beyond what could be inferred from their quantity. Most tweets followed the expected code of conduct for police while not breaking with the flexibility of the codes of conduct on Twitter. In effect, most often, the Norwegian emergency room tweets were not intentionally humorous. The tweets that gained attention were those in which the paradoxes became apparent, where the institutional codes of the police collided with the code of informal language on Twitter.

The tweets that used humor were often connected to domestic incidents, further drawing boundaries of codes by employing dramaturgical setups. In research on humor, this setup is explained using incongruency theory, describing humor that involves the cognitive processes of the listener, wherein the first part of an utterance creates an expectation that is disconfirmed by the second part. [36] In this case, the mentioning of a possible crisis creates suspense, which is further heightened for readers particularly aware of the possibilities of unwittingly breaching privacy in these cases: “The police have responded to a call concerning possible domestic violence and loud screams from an apartment in Økernveien.” This is then followed up with an unexpected ending creating comic release: “Turns out there was a cockroach in the apartment.” [37] Indeed, the police seldom tweet about domestic incidents, and about half of the cases identified in the data were where they produce funny situations.

These types of tweets destabilize the situation through possibly expressing the violation of a code for police conduct—is this common interest or a private matter?—and then attempts to re-establish control through disclosing the ‘actual’ situation, which is innocent. However, read through the media reception, this play with codes makes other codes emerge, such as the connotation between police and danger. Indeed, the suggestion that the use of humor is dangerous, rather than inappropriate, puts the tweets in a different light. The subtle difference between a typical: “Sofiesgt: we are on site concerning a loud domestic incidence. No one is seriously injured. 1 man is brought into custody” [38] and the humorous tweets quoted above highlights the difference between potential danger materialized through police-Twitter. While the informational tweets always address an event that connotes danger in the form of crime or accident, the negative public response to the humorous tweets address the potential danger of privacy breaches. In other words, inappropriate expression by the police on Twitter is expressed in the media as potential danger.

In 2018, the National Police Directorate released official guidelines for language and conduct on official police Twitter accounts. [39] On an organizational level, these guidelines mark a compromise, an attempt to settle the battle of codes and define the sphere of Norwegian police-Twitter. The analysis so far suggests that the need for guidelines was not actualized through the combined corpus of tweets, but is related to the phenomenon as it was decoded through the media reports. Although there were differences between the police districts, the main corpus of tweets express serious information, with the occasional poem or attempt at humor and jokes. The guidelines addressed the issue as it emerged from the coded environment of Twitter in intra-action with the media, not from the corpus of existing tweets. While the guidelines were effects of an idea emerging from the negotiation of semiotic codes governing police and social media, they were conditioned on the materialization of the police on Twitter. What mattered was that the police were on Twitter, not so much what they were doing there. Doing, in this sense, was connected to the practice of engaging on a specific digital platform, not to the mundane events dominating the total corpus of texts. In practice, the semiotic codes emerged from the coded landscape of the platform, from the very fact that the police were navigating the technological space.

The official guidelines standardized the emergency control room tweet as an institutional form of information. This standardization breaks with the codes established for what the police should be doing on Twitter emerging from the main body of research on police on social media. The guidelines do mention the importance of creating trust and safety so that the public are not afraid to make contact, but in general the document is encoded, through descriptions of how to write in a way that connotes institutional authority: be matter-of-fact, sober, considerate, polite, calm; exhibit control, competence, and expertise; and be cautious with the use of humor. [40] While some of these words to some degree connote a human-oriented practice, such as “considerate” and “polite,” none of the words connote engagement or encouragement. The general tone of the guidelines suggests that one should hold back, make well-informed judgements, and create an atmosphere of authority closely related to that for which the police as an institution strive: that is, already closely resembling the computationally created topics and the formulation of the bulk of close-read tweets.

Reading these guidelines through the tweets as corpus, what emerges is a police-phenomenon writing itself through life’s various mundane events, about which someone should be doing something. Read through the media reception, however, the guidelines illuminate that police-Twitter is not simply a police phenomenon. The phenomenon of police-Twitter is not created solely through the practice of tweeting, or through keeping in line with the codes of policing while tweeting. It is entangled with factors outside of the institution, computational codes, and digital devices that spread the tweets widely throughout society to recipients that decode differently, and who continuously alter the meaning of the phenomenon. So, the police are funny on twitter. And they are not. But never at the same time.

Additionally, while one segment of the traditional media regularly comments on and questions police use of Twitter, news media tend to use police-Twitter as informants; several texts from the media define the police as an unbiased source of information about ongoing events. Here the code of police authority emerges as the governing convention of interpretation. This creates an interesting paradox where the phrase “The police write X on Twitter” is encoded as a verification of truth, while in most any other case, “[X person] writes X on Twitter” is encoded as a personal opinion or statement about private affairs, not something establishing facts about external events. So even if the body of police-Twitter in Norway is stabilizing, gaining contours of guidelines and practical use, the codes that govern it are not settled. Using code as a tool to explore the phenomenon over time, we discover that it is stabilizing, but not stable: it is still emerging and productive, entangled within the continuous emergences of connective cultures.

8_Concluding Remarks

The Norwegian police have become encoded in a quite specific way on Twitter, with a set of codes produced through intra-action between institutional codes of conduct, the mundanity of operational policing, and the everydayness of Twitter as a sphere. Despite the occasional public criticism concerning issues such as breaches of privacy, police-Twitter is not contested in the contemporary Norwegian context. Instead, it is more of a constant reconfiguration of codes entangled with the general scrutiny of police as enforcers of power, rather than a question of whether they should be present on the platform. In addition to the paradoxes that emerge in public debate from time to time, the general population has come to use police-Twitter to orient themselves about mundane events in their vicinities: how is traffic flowing; why is that helicopter overhead; what is that noise, that smell? In this sense, the phenomenon exists across spheres that are commonly understood as separate—the police, traditional media, the social Internet, the public—creating a phenomenon that can be analyzed using known codes to explore the codes emerging from and continuously defining the borders and contours of that phenomenon: again, stabilizing but never stable.

Using code as a material-discursive concept in the analysis of connective cultures helps us go beyond the classical semiotic understanding of language as a meaning-making structure. It helps us see that technical aspects and artificial languages—where code is quite literally understood as technological mechanisms—are part of communication, and it takes on meaning-making functions. The overall structures that emerge through topic modeling are entanglements of semiotic and digital code that provide a look into the meaning-making taking place on a not necessarily conscious level, where discourses develop slowly over time. The detailed level of close reading addresses the more traditional semiotic understanding of meaning-making, yet, read through the structures provided by topic modeling, nuances appear that would not be present in a traditional semiotic reading of the text. And, vice versa, the semiotic reading attending to the contradicting formal and informal codes in police-Twitter makes the output form the topic modeling into something more than just a list of mundane events. Together these levels of code highlight the entanglements within which meaning-making emerges in connective cultures. Technologies take part in meaning-making in ways that materiality always does, but that becomes more tangible and more readily analyzable when communication also appears as digital data.

_How to Cite

Guro Flinterud. “Codes Colliding in Connective Cultures: The Emergence of the Norwegian Police Emergency Control Room Twitter.” On_Culture: The Open Journal for the Study of Culture 14 (2022). <https://doi.org/10.22029/oc.2022.1304>.

CC-By-4-0
CC-BY 4.0

_Endnotes

  • [1] Many thanks to the anonymous reviewers and the editorial team for insightful comments and suggestions, and to my project team Brita Bjørkelo, Jenny Maria Lundgaard and Johanne Yttri Dahl for valuable input on draft versions of this text.
  • [2] Peck, Andrew, “The Death of Doge: Institutional Appropriations of Internet Memes,” in Folklore and Social Media, eds. Andrew Peck and Trevor J. Blank (Logan: Utah State University Press, 2020), 83–107.
  • [3] Yuri Lotman, Universe of the Mind: A Semiotic Theory of Culture (Bloomington: Indiana University Press, 2001), 13.
  • [4] José Van Dijck, The Culture of Connectivity: A Critical History of Social Media (Oxford: Oxford University Press, 2013).
  • [5] Lotman, Universe of the Mind.
  • [6] Karen Barad, Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning (Durham: Duke University Press Books, 2007).
  • [7] Barad, Meeting the Universe Halfway, 33.
  • [8] “Police Online Patrols,” The Norwegian Police Service, accessed May 29, 2022, <https://www.politiet.no/en/rad/trygg-nettbruk/police-online-patrol/>.
  • [9] Jeremy Crump, “What Are the Police Doing on Twitter? Social Media, the Police and the Public,” Policy & Internet 3, no. 4 (2011): 1–27.
  • [10] Christopher D. O’Connor, “The Police on Twitter: Image Management, Community Building, and Implications for Policing in Canada,” Policing and Society 27, no. 8 (2017): 899–912; Maggie Dwyer, “Reimagining Police Engagement? Kenya National Police Service on Social Media,” Policing and Society 30, no. 7 (2020): 760–776.
  • [11] Karen Bullock, “The Police Use of Social Media: Transformation or Normalisation?,” Social Policy and Society 17, no. 2 (2018): 245–258; Stephan G. Grimmelikhuijsen and Albert J. Meijer, “Does Twitter Increase Perceived Police Legitimacy?,” Public Administration Review 75, no. 4 (2015): 598–607.
  • [12] For a an extensive survey of police use of social media, see James Walsh and Cristopher O’Connor, “Social Media and Policing: A Review of Recent Research,” Sociology Compass 13, no. 1 (2019): 1–14.
  • [13] Kristin Asdal and Helge Jordheim, “Texts on the Move: Textuality and Historicity Revisited,” History and Theory 57, no. 1 (2018): 56–74.
  • [14] Many thanks to Kaisa Korsak for assistance with the lemmatizer in the Norwegian spaCy-pipeline.
  • [15] André Baltz and Fredrik Norén, “Tematisk innehållsanalys med temamodellering [Thematic Content Analysis with Topic Modelling],” in Digitala metoder i humaniora och samhällsvetenskap [Digital Methods for the Humanities and Social Sciences], eds. Johan Jarlbrink and Ferdrik Norén (Lund: Studentlitteratur, 2021), 211–234.
  • [16] Jianhua Yin and Jianyong Wang, “A Dirichlet Multinomial Mixture Model-based Approach for Short Text Clustering,” Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (New York: Association for Computing Machinery, 2014), 233–242. Doi: 10.1145/2623330.2623715.
  • [17] Simon Lindgren, Data Theory: Interpretive Sociology and Computational Methods (Cambridge and Medford, MA: Polity, 2020).
  • [18] See for example Chris Anderson, “The End of Theory: The Data Deluge Makes the Scientific Method Obsolete,” Wired, accessed May 29, 2022, <https://www.wired.com/2008/06/pb-theory/>.
  • [19] danah boyd and Kate Crawford, “Critical Questions for Big Data: Provocations for a Cultural, Technological, and Scholarly Phenomenon,” Information, communication & society 15, no. 5 (2012), 662–679.
  • [20] Lindgren, Data Theory, 15.
  • [21] On affordances, see for instance danah boyd, “Social Network Sites as Networked Publics: Affordances, Dynamics and Implications,” in A Networked Self: Identity, Community, and Culture on Social Network Sites, ed. Zizi Papacharissi (New York: Routledge, 2011), 39–58.
  • [22] Tarleton Gillespie, Custodians of the Internet: Platforms, Content Moderation, and the Hidden Decisions That Shape Social Media (New Haven and London: Yale University Press, 2018).
  • [23] Egon Bittner, “Florence Nightingale in Pursuit of Willie Sutton: A Theory of the Police,” in Policing: Key Readings, ed. Tim Newburn (Cullompton: Willan Publishing, 2004), 150–172.
  • [24] Kristin Asdal, “What Is the Issue? The Transformative Capacity of Documents,” Distinktion: Journal of Social Theory 16, no. 1 (2015): 74–90.
  • [25] Jenny Maria Lundgaard, Nød og neppe. Fra anrop til beslutning i politiets operasjonssentral [From Calls to Decisions in the Police Emergency Control Room] (Oslo: Universitetsforlaget, 2021).
  • [26] Lundgaard, Nød og neppe.
  • [27] See for instance Eli Skogerbø and Hallvard Moe, “Twitter på tvers – koblinger mellom journalister og politikere [Twitter Across – Connections Between Journalists and Politicians],” Norsk medietidsskrift 22, no. 3 (2015): 1–21.
  • [28] Judith Butler, Bodies That Matter: On the Discursive Limits of ‘Sex’ (New York: Routledge, 1993), 9.
  • [29] Original tweets from dataset, “Ring 3 vestgående ved avkj. Ulvensplitten, trafikkuhell. Trafikale problemer, kødannelse langt tilbake. Oppdateres. Tore Solberg, Op.leder,” Twitter, September 29, 2011, <https://twitter.com/oslopolitiops/status/119317440973901824>; “Trafikkuhellet ryddet, fin flyt i trafikken igjen nå :-),” Twitter, September 29, 2011, <https://twitter.com/oslopolitiops/status/119323431681523712>.
  • [30] Many thanks to control room ethnographer Jenny Maria Lundgaard for this insightful suggestion.
  • [31] Original tweets from dataset, “Brann i bil ved Ammerudhellinga, slukket, men trolig påsatt. Ingen skadde. Søker i området etter mistenkte,” Twitter, October 5, 2011, <https://twitter.com/oslopolitiops/status/121726218709118976>; “E6 nordg. før Karihaigen. Kjedkollisjon. Politi er på vei,” Twitter, October 4, 2011, <https://twitter.com/oslopolitiops/status/121284260945592321>.
  • [32] Kristjan Molstad and Martin Skjæraasen, “Oslopolitiet skrev ‘Årets tweet’ [The Oslo-police Wrote ‘Tweet of the Year’],” in Aftenposten, February 6, 2013, <https://www.aftenposten.no/oslo/i/Bx07/oslopolitiet-skrev-aarets-tweet>.
  • [33] Original tweet from dataset, “Storo: Vi fikk melding om pågående husbråk med kvinneskrik. Da vi kom fram fant vi et sykepleier-vorspiel. Vi forlater stedet snart,” Twitter, January 11, 2013, <https://twitter.com/oslopolitiops/status/289840348237283328>.
  • [34] Mari Lund Wictorsen, “Det er farlig å være morsom på Twitter [It Is Dangerous to Be Funny on Twitter],” in Aftenposten, February 6, 2013, <https://www.aftenposten.no/oslo/i/V66l/det-er-farlig-aa-vaere-morsom-paa-twitter>.
  • [35] Andrew Goldsmith, “Disgracebook Policing: Social Media and the Rise of Police Indiscretion,” Policing and Society 25, no. 3 (2015): 249–267.
  • [36] See for example Ida Tolgensbakk, Partysvensker; GO HARD! En narratologisk studie av unge svneske arbeidsmigranters nærvær i Oslo [A Narratological Study of Young Swedish Migrant Workers in Norway], (PhD dissertation, Oslo: University of Oslo, 2015), 96.
  • [37] Original tweet from dataset, “Politiet rykket ut til melding om husbråk og høye skrik i en leilighet i Økernveien. Det viste seg å være en kakerlakk i leiligheten,” Twitter, October 28, 2013, <https://twitter.com/oslopolitiops/status/394922067314765824>.
  • [38] Original tweet from dataset, “Sofiesgt: Vi er på stedet i forbindelse med et kraftig husbråk. Ingen alvorlig skadet. 1 mann kjøres inn i arrest,” Twitter, November 24, 2013, <https://twitter.com/oslopolitiops/status/404433346425413632>.
  • [39] National Police Directorate, Språk og innhold: Twitter i operasjonssentralene (Oslo: National Police Directorate, 2018).
  • [40] National Police Directorate, Språk og innhold.