The elementary forms of digital communication

Although a tremendous amount of modern interaction is electronic, our understanding of everyday digital communications—including what they look like and how their properties vary by medium and relationship type—is still growing. In this paper, we examine digital exchange in two of its simplest forms: email and SMS. Specifically, our data consist of 2,004 messages provided by a diverse sample of college students, supplemented by in-depth interviews with their authors. These data were collected in 2010—a time when both mediums were widespread but devoid of most of their modern complexity. Based on these data, we make two contributions: First, we develop an empirically grounded typology of the basic properties of text-based digital communication; second, we document the distribution of these properties across five common relationship types. Respectively, these findings provide a starting point to understanding the substance of digital exchange in all its many forms and an empirical benchmark for comparison.


Introduction
Digital communication is now central to everyday life. 90% of American adults go online and 96% own a cellphone of some kind [1,2]. And though digital technology is used for a staggering variety of ends-from reading news to buying products, watching videos to booking travel -for many, its core functions are social. We email colleagues, text romantic partners, and connect with friends, family, and strangers using a diverse array of social media [3].
Nevertheless, scientific understanding of everyday digital interaction has lagged behind its prevalence. Activity records contain data on the patterning of behavior on certain platforms [4][5][6][7]; surveys and interviews shed light on who uses them, how, and with whom [8][9][10][11]; and an enormous interdisciplinary literature considers how mediation impacts social exchange. Yet aside from public content, the substance of naturalistic digital interaction remains challenging to document: first, because findings from one platform may not generalize to others; second, because changing features and interfaces create constantly moving targets; and third, because examining private communications entails obvious practical and ethical concerns.
To help address this gap, we draw on previously unused data consisting of over 2,000 text messages and emails sent by a diverse sample of college students and supplemented by indepth interviews with their authors-data that were collected a 12 years ago. Each of these decisions was carefully motivated. Email and short message service (SMS) messages are not only staples of everyday use, but exemplars of textual digital communication in its most basic technological and social form (i.e. alphanumeric, asynchronous, dyadic, and non-anonymous). Yet they also present variation in terms of ease (full keyboard vs. keypad), accessibility (computer vs. cell phone), and length (unlimited vs. 160 characters). College is a time when social networks may be especially diverse [12,13] and the students in our sample were among the first generation of "digital natives," or people born into a digital world [14]. And in 2010, both email and SMS were widely used by young adults [15,16], but still devoid of their modern-day embellishments. (For instance, our data predate the availability of emojis on standard U.S. operating systems, as well as many other features-including the ease of sending photos, videos, and animations-now integrated into smartphones.) In other words, our data feature naturalistic digital interaction, at a time when it was ubiquitous but still uncomplicated, among a sample highly comfortable with its use.
Using these data, we pursue two objectives. First, we develop a detailed, inductive typology of the properties of text-based exchange. This framework-an interpretive guidebook of sorts -provides a thorough, if necessarily inexhaustive starting point for future research interested in categorizing, organizing, and comparing digital interaction in all of its many forms. Although prior work has generated such typologies, their scope tends to be much more limited. For instance, Boneva et al. [17] describe three types of emails that sustain relationships; Thurlow [18] identifies nine functional orientations of SMS messages; and many other authors address specific communicative properties in detail (discussed below). Our aim is at once to replicate, synthesize, and extend this research.
Second, we document the distribution of these properties across five common relationship types (peers, friends, family, authority figures, and romantic partners) and the two mediums (SMS and email). Beyond their utility as interesting and (to our knowledge) novel findings in their own right, these patterns provide an empirical benchmark for a variety of possible replications and comparisons: with other populations, relationships, and communication technologies and to understand how email and text message use have changed over time. Importantly-given the vast, variegated universe comprising modern digital life and the sheer volume of studies examining its many corners-we also hope they will help illuminate what interactional dynamics may be unique to certain platforms or potentially generic to text-based exchange. In other words, we seek to exploit the simplicity of our data to understand modern digital communication in all of its complexity-a counterintuitive but familiar strategy to most sociologists [19]. This paper is organized as follows: We begin with a targeted review of relevant research, emphasizing how our approach compares to past scholarship on digital exchange. Next, we discuss our data and methods. Results are presented in two sections: one documenting the multitude of ways email and SMS are used and one demonstrating how usage varies by medium and relationship type. We then consider the larger implications of our results and conclude by returning to the modern digital context.
Research on email shares many of the same emphases. These include studies of how email is used to maintain ties [17,73,74]-especially over long distances, as with migrants [75]and examinations of its linguistic features [76,77] or general content and purpose [78]. Earlier work focused on email in the workplace [79]. Other notable strands of research compare email communication with phone calls or voicemails [80,81]; examine gender differences in message content, recipients, and perceptions of usefulness [17,82,83]; and use patterns of email exchange to learn how social networks form and evolve [5,84,85].

Limitations
Although certainly not true of all studies, we offer the following broad comments on the preceding body of work. First, much prior research takes textual communications at face value without exploring their subjective meaning to authors and audiences. On the other hand, relying solely on self-report presents an equally incomplete portrait of human behavior. Comparatively few studies [38,58,78,86] examine the content people produce together with their understandings of why they produced it.
Second, past research on the content of texts and emails commonly focuses on one aspect of messages, such as their "semiotic tactics" [86], degree of cultural fit [87], nonverbal cues [88], or use of relational management strategies [73] rather than exploring messages more holistically and considering substantive, stylistic, and functional dimensions in tandem.
Third, due to data access and privacy concerns-and notwithstanding a number of important exceptions [18,58,73,88]-it is still difficult to examine naturalistic exchanges, i.e. the everyday messages people send and receive outside the confines of the laboratory or workplace. Fourth and relatedly, past research frequently focuses on a particular kind of interaction [89], such as among coworkers [31,90], romantic partners [91,92], or strangers [86] or else includes a broad set of relationships but does not systematically distinguish among them [38,58,62]. Respectively, these tendencies obscure the breadth and versatility of modern digital communication and ignore the basic insight that a core determinant of any performance is the audience to whom it is tailored [36,73,93].
As Baym [94:59] suggests, "Instead of asking what mediation does to communication, we can also ask what people do with mediated communication" (italics in original). We pursue this guidance in the remainder of this paper. Specifically, we shed light on the digital encounters of college students, using 1) a mixed-methods investigation of 2) a variety of features of 3) naturalistic text messages and emails, including 4) systematic comparisons by relationship type-identifying the distinct opportunities for connection afforded by text-based digital technology.

Materials and methods
Our methods consist of content analysis and in-depth interviews. In late 2010 and early 2011, we requested a total of 150 text messages and emails from each of 15 research subjects, yielding 2,004 distinct communications (some subjects could not provide some types of messages; see below). We coded these communications for a variety of properties and interviewed all subjects about their perceptions and usage patterns of SMS and email. Our study was approved by the Harvard University Committee on the Use of Human Subjects (where both authors were affiliated at the time) and informed written consent was obtained from all participants. Here, we describe the details and motivation of each step in our research design. and mainstream (as opposed to merely work-related) email use and so helped establish many of the nuanced social rules surrounding these technologies. To identify potential subjects, we began by contacting acquaintances of the first author (who was then a senior in college), who in turn connected us with friends they believed would be interested in participating. Potential subjects received a series of three emails prior to participation (see S1 File). We sent the first email to 23 contacts, the second to 19, and the third to 17-two of whom dropped outamounting to a final sample of 15 or a response rate of 65%.
Given the nature of our study, we took for granted that any type of random sampling would be impossible (insofar as very few people would be willing to share such private material with strangers and set aside time to be interviewed). We hoped that by targeting contacts removed by two degrees, we would harness the trust benefits of a shared acquaintance [95] while slightly reducing the biases of convenience sampling. Still, we attempted to recruit a sample that was as socio-demographically diverse as possible. Participants represented a range of racial identities (including black, white, Hispanic, and Asian) and socio-economic backgrounds (from lowermiddle to upper class) and grew up in a variety of American cities and towns. They attended many different four-year colleges and had an average age of 21. Eight were male and seven were female.

Text and email collection
We requested 55 emails and 95 text messages from each participant, sent to five categories of recipients: authority figures, close friends (same-and opposite-sex), immediate family members, romantic partners, and peers who were not close friends (see overview in Table 1). These categories build on prior research [17,73,89,96,97] and represent a diversity of expected age differences, intimacy levels, and power dynamics. Securing a random sample of messages seemed as unlikely as a random sample of respondents. In our best effort to maximize data representativeness while following IRB guidelines, we clearly communicated our research goals and asked generally for "a sample" of messages. (While participants were repeatedly reassured of the confidentiality of their data, they were also clearly instructed not to provide messages that they were uncomfortable sharing; see S1 File).
Though we requested messages from the past month, in some cases older communications were unavoidable (for example, many students only interacted with a boss during the summer). Participants also reported difficulty finding 20 communications to non-friend peers, so we lowered this requirement to ten. Subjects typically copied and pasted emails into a formatted checklist we provided, but submitted text messages in a variety of ways depending on the kind of phone they used (e.g. some with iPhones uploaded conversations to their computer and sent us PDFs; some with Blackberries emailed themselves conversations and forwarded excerpts; others manually transcribed their own messages). We permitted emails addressed to multiple recipients but not to a formal listserv. (Texting multiple recipients was not yet widespread.) If subjects accidentally submitted messages they had received as well, we immediately deleted them; we examined outgoing messages only.

Inductive coding
We followed a constructivist grounded theory approach [98] to interpret the messages through textual analysis. We let concepts emerge from the data [99] and first used open coding to develop codes inductively [100]. We then used focused coding to distill these codes in tandem with data analysis to ultimately develop four data-driven code categories: A) the intended purpose of the communication; B) the substantive and C) discursive techniques the author used; and D) the social function of the message. Respectively, these categories address the following questions: A) What is the reason for the communication, or what is the message doing? How does the author communicate that purpose, in terms of B) the information she includes and C) the way she expresses it? And D) what is the plausible impact of the communication on the relationship between its sender and recipient? To account for differences in message length and enable meaningful reporting of the distribution of codes across communications, we applied codes to entire messages as opposed to particular words or sentences. The same message could receive multiple codes within and across categories. We performed all coding using ATLAS.ti.

Interviews
The first author also conducted semi-structured, open-ended interviews with each of the 15 participants. These lasted approximately half an hour to 45 minutes and took place either in the home of the respondent or at a neutral location, like a café. Most interviews were conducted in Boston, MA, though some took place in Middlebury, VT and Chicago, IL-each of which represented the city where the respondent either attended college or grew up (and so returned for breaks). All interviews were digitally recorded and later transcribed. The interviews were intended to gauge subjects' general impressions of SMS and email exchange and understand their subjective experiences communicating through these mediated forms. Specifically, respondents were asked how frequently they used SMS and email, to whom they sent messages, and in what situations they preferred each medium. They were also asked about their perceptions of authenticity and privacy in digital exchange; their beliefs about and experiences with collaboration in message composition; their motivations for the creation of different fronts and techniques employed to do so; the impact of spatial distance from the recipient on presentation of self and teamwork; and norms surrounding digital exchange, such as meanings ascribed to response times, expectations regarding situation-dependent use of SMS versus email, and the relative formality of each. The interview guide and consent form are presented as S1 File.

The properties of text-based digital communication
We present our results in two sections. The current section documents the codes that emerged from our data, organized by the four communication properties (code families) identified above. For each code, we explain its features and provide ideal-typical illustrations. The following section of results then assesses the empirical distribution of these codes by medium and relationship type.
For reference, a list of all major codes is presented in Table 2. Messages are copied verbatim (except for names, which have been replaced with pseudonyms throughout), so apparent typos reflect the original message. When helpful, we cite interview data supporting claims about intent. Because our aim is to generate a broadly applicable framework (rather than examine how these features are patterned, as in the next section), we present SMS and email together. Codes are introduced in each sub-section in rough order of increasing intimacy-a central distinction that arose inductively, as in past research [18].

Intended purpose
Email and SMS communications tend to serve one of seven purposes. Young adults use text messages and email for the most impersonal relays of information to the most private confessions, demonstrating the broad role of these communications in everyday interaction. Thurlow's study of SMS [18] produced similar categories; we identified these purposes in SMS and email alike.
Practical information. These communications deal with the exchange of pragmatic details, specific requests for information, or answers to such requests. Reading a text or email in this category, it may be difficult to discern the nature of the relationship. These messages most frequently deal with events or actions already in progress or with plans for or questions about the future, i.e. "logistics." Practical arrangement. These messages also exchange functional information. However, while practical information deals with established or projected facts, practical arrangements are not yet fixed-they require coordination between sender and recipient. They are thus "arrangements" in the traditional sense: agreement on the specifics of future plans. They ask or answer the questions "Where?" (IP.7), "When?" (IP.8), or both (IP.9). Non-social arrangement. This category (along with the more intimate "social arrangement," below) also refers to arrangements but surpasses logistics. This includes the initial expression of a desire to meet, discussion regarding the purpose of a gathering, and any nonrecreational plan-making; arrangements in this category serve a functional purpose in the lives of their senders. IP.10. I wanted to talk to you about potentially interviewing you for my thesis. (F email to authority) IP.11. Do you think we could set up a time to meet briefly next week to go over what I will have missed? (M email to authority) Both IP.10 and IP.11 are about plans for formal meetings. In many instances (e.g. IP.11), this purpose is expressed in conjunction with a practical arrangement. This may occur between individuals who do not interact frequently and thus engage in economical exchange, meaning they include multiple elements of plan-making to eliminate unnecessary dialogue.
Corrective face-work. These communications attempt to repair a lapse in performance on the part of the sender-either by conveying explicit remorse (IP.14) and/or attempting to absolve the sender of any accountability for a past, present, or predicted behavior (IP.12, IP.13).
IP.12. sorry didnt respond earlier absolute crazay week here (F text to peer)

PLOS ONE
The elementary forms of digital communication IP.13. I really didn't want to just email you like this. . . I have tried to stop by your office several times this week, but you haven't been there. (F email to authority) IP.14. Hey, so sorry to not have texted earlier (M text to friend) Corrective face-work can occur on its own or with other purposes. When recoveries were embedded in messages with other purposes, closer inspection often suggested that these purposes were superficial vehicles for the deeper objective of corrective face-work.
IP.15. Hi Professor Cohen, I hope that the week has been treating you well! I am so sorry for not stopping by today as I had said I would. I ended up shopping a seminar last minute that just let out [italics added]. I will be taking the WGS class that I was telling you about, as well as an English class this semester.
I am off to meet Sam now-once again thanks so much for putting me in touch with her! Talk to you soon, Stephanie (F email to authority) IP.15, for example, relays non-practical information (see below) by keeping the participant's professor updated on her plans for the semester. However, recovery was the driving motivation behind this communication; Stephanie included the other details to obfuscate her aim of corrective face work "because otherwise it would seem like I was just writing to cover my ass" (Interview 6).
Non-practical information. Communications in this category convey inessential facts or details. These range from gossip to jokes to personal details irrelevant to the recipient (like Stephanie's class selection, above). These messages do not have a practical objective and typically contain jovial, sociable qualities indicating the exchange of information for the mere pleasure of exchange.

Substantive techniques
While some scholars have argued that basic textual communication lacks social and emotional cues, our data revealed quite the contrary. In emails and text messages, implicit information is paralinguistic, rather than non-verbal (as in face-to-face interaction). In the absence of facial and tonal cues, body language, and physical props, individuals compensate with purposeful choices in writing style, composition, and expression [27]. This sub-section deals with four substantive techniques regarding what kind of information is communicated. The next, on discursive techniques, discusses how.
Qualify relationship status. Codes in this category situate the relationship between sender and recipient. While not all communications-particularly those that were part of an ongoing exchange-contained such statements, those that did took one of three forms: First, in messages establishing contact with a previously unknown person, participants engaged in an introduction. Unlike some polite face-to-face introductions that occur when individuals happen to find themselves together, text and email introductions are motivated by specific functional purposes such as a job hunt or class enrollment. (Otherwise, the message wouldn't have been sent).

ST.28. My name is Penelope Miller and I am a junior undergraduate studying computer science at Harvard University. (F email to authority)
Second, when communicants already know one another but are not in frequent contact, a reference to a past encounter, either digital or face-to-face, may be used.
ST.29. To put a face to a name, we briefly met during a book signing at Harvard in October, when you came to speak about Half the Sky [book title] (although considering the number of people who were there, I'm not sure that this will trigger much of anything!) (F email to authority) Third, participants made reference to a future encounter. Such messages contain explicit expressions of hope for continuing the relationship and/or interest in the recipient's response.
ST.3o. Looking forward to hearing back from you. (M email to authority)

ST.31. Can't WAIT to see you on Friday! (M text to friend)
Each of these techniques situates the communication for the recipient in terms of a (nascent, ongoing, or projected) narrative with its sender.
Manner of address. These codes refer to how subjects addressed recipients. Often, there was no address at all; but if there was, its nature helped establish the tone of the message.
If a participant addressed a recipient by title (e.g. Mr., Professor, Dr.), we considered it a formal address. Subjects used "Hi," "Hello," and "Dear" interchangeably and, on occasion, excluded a greeting and merely began with the recipient's name. In all such cases, the communication structure was formal as well, including a traditional sign off such as "Best," "Sincerely," or "Thank you." ST.32. Dear Professor Rosen, . . .

Best, Christine (F email to authority)
Unlike formal addresses, nicknames or terms of endearment were often present throughout messages, acting as tokens of intimacy rather than respectful salutations [58]. Although nicknames (ST.33, ST.34) are individualized while terms of endearment (ST.35, ST.36) may be generic, we group them together given their functional equivalence. In some relationships, otherwise insulting names take on positive meaning (ST.37); in the privacy of light-spirited digital rapport, they are markers of closeness and affection (by highlighting the fact that the relationship is close enough to permit such words in jest).

ST.37. hussy! I miss you slut (F text to friend)
In sum, distinct forms of address convey the sender's perception of their relationship with the recipient to the recipient: one establishing social distance and one fostering closeness. One participant categorized all digital interactions based solely on this distinction, asserting "There's two kinds of things: dear so and so and then you sign your name. And then there are the ones that are one line or whatever without the formality" (Interview 6).
Emotional expression. Without tonal cues, facial expressions, and body language, textbased digital interactions rely heavily on the explicit verbal expression of information taken for granted in person. Subjects expressed this information through purposive diction and overt articulation of feelings [33,101]. These tools are among the most important for digital interactants because, more than any other substantive technique, they influence how cointeractants interpret their words.
Because we identified so many distinct emotional expressions-as with subject matter, below-we summarize them in Table 3. This category encompassed the following 12 codes: deference; personal detail; gratitude; relatability; well wishes; enthusiasm; empathy; check-in; support; for you. . .; "love"; and romantic expression.
Subject matter. Unlike emotional expressions, which involve explicitly voicing feelings, subject matter refers to the focal content of a message-not unlike "what people talk about" offline [102]. This category encompassed the following 11 codes, summarized in Table 4: confirmation or agreement; apology; excuse; humor; banter; micro-coordination; link; gossip; collusion and secrets; sexual matter; and derision.

Discursive techniques
As scholars have long noted, the lack of sound or graphical content even in exclusively textbased CMC has not prevented users from finding other creative ways to increase its richness and illocutionary force [27,58,103]. Discursive techniques are paralinguistic elements of textual communications used to modify meaning and convey emotion [101,104]. Traditionally Table 3. Types of emotional expression.

Code Explanation Examples
Deference Encompasses a variety of expressions (e.g. respect, a relaxed personality, indifference, ambivalence); enables the recipient to feel comfortable and make decisions favorable to her own interests. It is also, consequently, a face-saving mechanism: By deferring to the recipient, the sender ensures she will not make the wrong choice.
ST. 38 For you. . . Instances when subjects send a piece of information-whether an anecdote, web link, or unprompted favor-they think the recipient will enjoy. Sometimes senders directly state that they are "thinking of you," but usually this is implicitly evidenced by how thoughtful and personal is the "gift." ST. 55

PLOS ONE
The elementary forms of digital communication

Collusion and secrets
When interactants surreptitiously convey information they do not want others to know; digital communication is thus used to both confide in cointeractants and ensure privacy in these exchanges. We group collusion and secrets together because information frequently straddles the line between disclosure of confidential information (secrets) and putting that information to use for an intended clandestine purpose (collusion). associated with personal style, they can also indicate a purposeful change in register or tone. Some, however, are borne of mere convenience [18]. Discursive techniques almost always represent informality; exemplary of this is one participant's confession that if anyone besides a friend or close family member saw her typical digital writing style, "they would think I'm not good at English" (Interview 7). We identified five such techniques in our data. Abbreviations. Abbreviations are common in text messages and emails-as on the internet at large [105]. They convey a sense of casualness, depend on the subject's writing style rather than an intention to alter meaning, may be specific to a particular in-group, and result most likely out of ease [18,58]. They include contractions ("wk" for "week"), acronyms ("omg" for "oh my god") and shortenings ("bro" for "brother"). Subjects emphasized the informality of abbreviations, as in the following quotation: "If you're in an academic setting, you don't want them to think you're dumb, so obviously you proofread and use real words like a grown up" (Interview 2). They are interpreted as a sign of "speed" (Interview 13) and indicate the message is "unofficial" (Interview 6).
Intentional misspelling. Participants use misspellings [58] to convey personal registers and make the tone of a message more representative of their own speaking. The presence of such choices is user-dependent and, unlike other discursive techniques, they are more a tool for personalization than for emphasizing certain details or expressing specific emotions. Examples include eliminating letters ("hav" for "have"), atypical spellings ("shud" for "should"), and accent representation ("sumfin" for "something").
DT. 85 M text to friend) a Many participants conveyed a preference for expressing both apologies and excuses over SMS or email, rather than face-to-face-a finding we return to in the discussion. b Indeed, many subjects felt that digital exchanges could be just as much of a conversation as face-to-face exchanges. One participant said, "I don't see the differenceyou're going back and forth the same and you're just talking in written form" (Interview 11). Participants overwhelmingly perceived continuous, rapid-fire exchanges like ST.66 as "basically the same thing" as in-person encounters, which accounts for their apparent triviality and geniality. c One participant articulated the distinction between links that do and do not require responses. He explained, "If I sent an email to my sister of a stupid YouTube video. . . I know that she'll laugh when she reads it, I don't need to hear that via email" (Interview 8). The purpose of these messages is typically to entertain the recipient and link-senders can feel they have accomplished this without explicit confirmation-presumably because they feel sufficiently confident of the recipient's tastes. Emoticons or symbols. Emoticons are facial representations created by sequences of punctuation, such as ":)" (a smile), and symbols are sequences of letters and numbers that represent a gesture or emotion, such as "xoxo" (hugs and kisses). A precursor to emojis, they are generally employed to communicate, clarify, or emphasize sentiment, often by mimicking physical expressions or objects [103,[108][109][110][111]

Social functions
This category of codes addresses the micro-consequences of digital communications, or the impact of messages on the relationship between the people exchanging them. Here, elements beyond the overt contents of communications are as important to their classification as the more obvious features presented above. Though we did not solicit data on hostile relationships (e.g. "enemies"), all messages provided by our respondents appeared to augment, rather than weaken, social connections. Because these functions are accomplished by a wide array of communication types-and to underscore this point-we exclusively reference earlier examples and introduce no new data at this stage.
Establish relationship. Digital messages frequently serve as inaugural communications between unfamiliar parties. Because of the privacy and flexible time horizon in which they can be crafted, they are often a more comfortable way to begin contact compared to the pressure of real-time introductions. ST.28 illustrates such an introductory email with a potential employer, whom the participant said she would have felt too "awkward" and "nervous" to approach in person or on the phone (Interview 2). Importantly, email thus eliminates the necessity of a physical meeting-which can be inconvenient or impossible-to begin a relationship. For instance, the participant in ST.42 is inquiring about potential opportunities for graduate school. She commented that it would have been much more difficult to contact professors at various institutions had email not been an option (Interview 12).
Maintain face. A common function of messages is to help senders preserve their selfimage [58]. However, this is not limited to messages whose purpose is corrective face-work (e.g. IP. [12][13][14][15]. In all digital interactions, participants make discursive and substantive choices in keeping with their role. Consequently, many messages not only serve their intended purpose, but also play a strong role in face maintenance. For example, IP.4, which employs a range of discursive techniques and focuses on a social subject-in this case, buying alcohol for a gathering-preserves the subject's already established "friend front," i.e. one that is lively and social. Conversely, messages like ST.38, ST.48, and ST.50 preserve a different kind of image by revealing their sender to be respectful. Maintain ties. For people who cannot see each other frequently, emails and text messages are an important part of staying in touch. In ST.24, for example, the participant reaches out to her father, maintaining a close connection despite being physically distant. Digital communications also allow subjects to acknowledge important people in their lives without a large time commitment. For example, the playful tone and simple acknowledgment "I miss you" in ST.37 convey that the sender thinks about the recipient and wishes she were nearby. Even the simplest message-a link, for example (IP.17, ST.70), or a quick well wish (ST.46)-indicates to the recipient that the relationship is important. Communications like this can be sent to anyone, at virtually any time, making ties that might otherwise be forgotten or neglected (e.g. a grade school friend or a great aunt) easy to sustain.
Increase solidarity. Digital communication increases solidarity in multiple ways. On one hand, it facilitates high rates of interpersonal contact, enhancing familiarity and cohesion. On the other hand, it allows individuals to discover and emphasize similarities with each other that may not be noticed or acknowledged in person. For example, in ST.44, the participant expresses to her professor interest in the Na+/Ca2+ exchanger. She goes on to explain this interest in detail-something she does not have the opportunity to do during regular in-class interactions (Interview 1). Banter, gossip, derision, collusion and secrets, and expressions of confirmation or agreement are other common means by which individuals feel close and affirm like-mindedness-occasionally by exclusion (see below).
Enable mutual activity. Text messages, and particularly emails because of their unlimited length, allow individuals who are not physically copresent to engage in mutual activity; beyond passively reporting on what one is doing or experiencing [89], others can actually be involved. In ST.39, for instance, the cointeractants are planning a party. Another example is collaboration in online shopping: By sending a link to a shopping website, individuals can seek the advice of friends before making a purchase (ST.72). Indeed, digital technology facilitates such exchanges even when they are unnecessary; of emails like ST.72, one subject said, "It's an excuse to talk to someone if you didn't have other particular things to say" (Interview 12). While mutual activity clearly contributes to solidarity, it also creates the special feelings of unity and accomplishment that accompany task completion.
Emphasize team membership. The basic act of digitally communicating with multiple recipients serves an important function for teams. Group messages go above and beyond enhancing solidarity by explicitly distinguishing insiders from outsiders-boundaries unlikely to be so demarcated in person. By sending group messages, the inner circle acknowledges to themselves their own insider status. Furthermore, such messages create a safe, "virtual backstage area" where team members can speak freely and act without concern for any audience. For example, both ST.71 and IP.21 were sent to multiple recipients. In ST.71, the author mocks his cousin; in IP.21, the author confesses she is not having a good time at the bar but feels a social obligation to stay. Beyond group messages, gossip and derision also mark team boundaries in less obvious ways; by alienating those being discussed, interactants bring themselves closer together and confirm membership to the same 'side.' ST.71 is an example of this, as are ST.74 and ST.81.
Confirm intimacy. Finally, individuals use digital interaction to affirm the closeness of relationships and express affection. In ST.55, this expression is implicit: The sender goes out of her way to find dresses she thinks the recipient would like. In ST.58 and ST.59, it is explicit: The sender openly professes feelings of love or sexual attraction. The ease and accessibility of SMS and email are again important because they allow users to quickly send short messages of love, even in times apart (as in ST.56). That these messages are written proves especially crucial. A typical example of this is ST.57; the sender ends a text message coordinating a plan to "talk soon" by including "Much love xx." By compensating for the lack of corporal and tonal signifiers (if these friends had been in person, perhaps they would have simply hugged goodbye), individuals may end up explicitly highlighting emotions that would otherwise go unmentioned-thereby heightening affirmations of intimacy [26].

The distribution of properties by relationship type and medium
In the previous section, we drew upon respondents' SMS and email communications and explanations thereof to develop a set of tools to distinguish, classify, and analyze their exchange. Employing these tools, we now examine the characteristic features of communications across five types of relationships and both mediums. We begin each sub-section with a summary description; present detailed statistics on the prevalence of specific codes; and conclude with an exemplary email and text message of relationships of that type to provide a holistic, qualitative illustration of the patterns we have described. The complete distribution of codes by medium and relationship type is presented in Table 5.

Authority figures
The most common authority figures with whom respondents communicated were professors. Messages to authority figures are characterized by formality, exchanging practical information, and maintaining relationships, particularly manifested in corrective face-work. They primarily serve to heighten solidarity and maintain face. Because the informality of text messages is incompatible with the front that students are here trying to maintain, almost all data in this sub-section come from emails.
Intended purpose. Of the 181 total communications with authority figures (150 emails and 31 texts), the most frequent purposes were to exchange practical information (44% of emails, 74.2% of texts), engage in corrective face-work (28.7%, 35.5%), and make a practical arrangement (20.7%, 9.7%). Many face-work emails refer to mistakes made in person, such as: "I'm sorry for not stopping by today, as I had said I would." (Excerpt from IP.15) Email is a recovery tool with small costs: It takes little time and energy to craft a message that could recover face. Indeed, one wonders whether the student may have been more likely to visit the professor's office if apologizing for not doing so was not so easy.
Only four participants sent texts to authority figures. In all instances, the recipients were subjects' bosses at on-campus jobs and young adults themselves-suggesting a potentially important interaction between relationship type, age differences, and digital medium.
Substantive techniques. Less than half (42%) of emails to authority figures qualify the relationship. The majority of these reference a future (28%) or past (8.7%) interaction, and very few (5.3%) make introductions. This indicates subjects tended to contact authority figures with whom they already had a relationship-though not necessarily a close one. In fact, in 141 of these emails (94%), subjects addressed the recipient formally. 120 (80%) express gratitude and 43 (28.7%) express deference. Other common emotional expressions are well wishes (27.3%) and relatability (18%), reflecting a desire to strengthen the social tie. The most Subjects used text-based digital communication more to preserve ties than to strengthen them. Example.
Dear Professor Terrell, I hope you had a good weekend! I am writing regarding the most recent paper, which is attached. I am really sorry that it is late; I was dealing with a sudden personal crisis and just couldn't finish it by the deadline yesterday. Please do not take this tardiness to be any indication of my commitment to the class, which I am enjoying thoroughly and find really interesting.
Looking forward to lecture tomorrow,

Melanie (F email to authority)
[All subsequent sub-sections provide illustrative emails and text messages. The infrequency of texts to authority figures makes all such messages atypical and it would be misleading to suggest one as a prototype.]

Peers
Though similarly polite, messages to peers contain a wider array of features than those to authorities and have more personal character. They often contain practical information coupled with attempts to connect with the recipient, such as through gratitude, relatability, or confirmation. Many use corrective face-work-typically excuses rather than apologies. Unlike with authorities, text messages are employed freely. They serve to maintain face and increase solidarity. Intended purpose. To peers, participants were most likely to send emails about practical information (66.7%), while texts include both non-practical (38.7%) and practical (32%) information. Interestingly, 26.7% of emails contain corrective face-work, but only 4% of texts. This reinforces the main difference we observed between email and SMS: Because emails are longer, more susceptible to formal organization, and easier to compose because of the full-sized keyboard, they tend to be used in an "official capacity" (Interview 4).
Substantive techniques. When engaging peers, subjects tend not to contextualize the communication. The most frequent form of qualification occurs in texts that reference a future communication (8%). The most commonly expressed emotions are gratitude (17.3% for email, 16% for text), relatability (13.3%, 10.7%), enthusiasm (13.3%, 12%), and well wishes (13.3%, 6.7%). Though choice of medium depends on message purpose-email for administrative and SMS for social tasks-both convey a similar range of emotions. The information most frequently contained in texts is confirmation or agreement (16%), while in emails it is excuses (17.3%). Second most frequent in SMS is micro-coordination (5.3%).
Discursive techniques. Discursive techniques are used slightly more often with peers than with authority figures, particularly in emails. Only 3 texts (4%) contain any discursive techniques, while 23 emails (30.7%) do. In emails, the most common techniques are extra letters (9.3%), emoticons or symbols (8%), and intentional misspellings (6.7%). Still, these are relatively low figures in comparison with messages to family, friends, and romantic partners (below).
Social functions. The most frequent functions of communications with peers are maintaining face (29.3% of emails, 9.3% of texts) and increasing solidarity (10.7%, 16%). That these are the two most common functions of messages to authority figures as well points to the similar role text-based communication serves in these very different types of relationships. Examples.

Family
Communications to family members are characterized by the exchange of information, both practical and non-practical. They are casual, cheerful, and thoughtful and convey extreme comfort with their recipient. Subjects primarily use text messages to micro-coordinate and confirm or agree and emails to joke and send links. Such communications maintain ties, increase solidarity, and confirm intimacy. Intended purpose. The most frequent use of digital communications among family is relaying practical information (40.7% of emails, 32.6% of texts), followed by relaying non-practical information (32.7%, 36.4%) and relationship work (9.3%, 25.8%). That these three purposes collectively make up almost all familial communications (88.3%) is notable. The low incidence of arrangement messages may not be generalizable; because our sample is predominantly composed of students who attend college away from home, it likely reflects their distance from families rather than typical communication habits.
Substantive techniques. Individuals are less likely to qualify their relationship status or use any form of address with family than with anyone else. Because we requested communications with immediate family members only, these data reflect recipients with whom subjects most likely have frequent and longstanding interactions. This, combined with a high level of intimacy, means contextualization is generally unnecessary. The most frequently expressed emotions are gratitude (17.3% of emails, 7.6% of texts), enthusiasm (12.7%, 4.5%), and love (21.3%, 1.5%). In all, there are more emotional expressions in emails-likely due to the length restrictions of SMS-but more check-ins in text messages-because such messages are short and intended to be immediately received.
The type of information exchanged with family is more dependent on medium than the two previous relationship categories. In email, most frequent are humor (20.7%) and links (17.3%)-and then, with a substantial drop, apologies (6.7%) and confirmation or agreement (6%). In texts, most common are confirmation or agreement (13.6%), micro-coordination (12.9%), and humor (9.1%). Humor is used with family more than any other relationship besides friendships, reflecting an overriding geniality to these communications.
Discursive techniques. Participants employ more discursive techniques with family than with peers or authority figures. This is representative of their generally increased emotionality compared to the previous, more socially distant categories of recipients. Using extra letters is the most widely applied technique, occurring in 7.3% of emails and 6.1% of texts.
Social functions. Communications with family serve two general functions: maintaining ties (34.7% of emails, 15.2% of texts) and increasing solidarity (32%, 15.2%). Emails also confirm intimacy (16%) and engage cointeractants in mutual activity (12.7%). This certainly relates to the high level of non-practical information families exchange; links, too, provide subject matter for jocular discussion. By keeping family members tuned in to their daily lives and sustaining casual dialogue, subjects accomplish emotional closeness despite physical distance. Examples.
hayyy im heere where are you? (F text to family)

Friends
Respondents text and email friends persistently throughout the day; these exchanges are arguably as integral to friendships as face-to-face interaction. Unlike previous categories of relationships, these messages primarily concern recreational plans and friendship work. This is accomplished through high rates of emotional expression, including support, check-ins, gratitude, love, and frequent use of humor. Banter and link-sending are common. Between friends, digital interaction heightens solidarity, affirms team membership, and confirms intimacy. Though we separately requested communications to same-and opposite-sex friends, there were few systematic differences in the features of each; we therefore treat them here as one category.
Intended purpose. Communications among friends are highly intimate. Practical information is still prevalent (22.3% of emails, 9.3% of texts), but social arrangements (11.3%, 10.2%) and relationship work (72%, 46.8%) are more common here than for any other relationship type. The heavy reliance on these digital mediums for friendship work shows they have been fully integrated into the fabric of daily communications among strong ties. Moreover, the difference in non-practical information exchange between friends (21.3%, 16.9%) and family (32.7%, 36.4%) suggests texts and emails are more integral to the day-to-day execution of friendships-as opposed to keeping family informed about one's life.
Discursive techniques. Of all types of relationships, individuals are most likely to use discursive techniques in messages to friends. Most frequent are extra letters (9.7% of emails, 4.3% of texts) and intentional misspellings (9.7%, 3.2%)-again pointing to the casualness of these exchanges and the informal tone participants create. We attribute the disparity in these techniques between email and SMS to interface: Most are easier to accomplish on a keyboard.
Social functions. Through digital messages, friends confirm intimacy (15.3% of emails, 13.1% of texts), emphasize team membership (22.7%, 14.7%), and increase solidarity (23%, 15.1%)-functions that deepen socio-emotional bonds. Gossip and derision validate the divide between "us" and "them." Confiding in a friend expresses trust and further strengthens the relationship through mutual preservation of a secret. Furthermore, by showing recipients how they are like-minded with the sender, high rates of confirmation continuously heighten solidarity. Examples.
hiiiii! how are you lover?? we should talky soon miss you kiss you (F text to friend) haha true good point, but i was dying when i read it [www.whenparentstext.com]. and the eighth down. i've got to stop looking at this site and write these goddamn papers, shittttttt (M email to friend)

Romantic partners
Similar to friends, partners use text messages and email to stay in touch throughout the day, engaging in sustained romantic relationship work and banter rather than achieving practical ends. Communications are characterized by overt expressions of affection and a cheerful tone; they are positive and warmhearted. Partners seldom employ discursive techniques. Rather, messages check-in, convey support, and often contain humor. They increase solidarity and confirm intimacy. As with some other categories of relationships, our data on romantic partners are incomplete. Of the 15 total participants, only five females and five males had someone they would consider a partner. While all ten provided the full ten SMS messages we requested, only eight (four women and four men) exchanged emails with their partner, and one of these women had sent only three. As the male with no romantic emails explained, "I'm always just texting with girls. I really only send emails if I have to; emailing a girl would be weird. The only girl I email is my mom" (Interview 3).
Intended purpose. The primary purpose of messages between partners is relationship work (42.5% of emails, 49% of texts), followed by relaying non-practical information (35.6%, 17%). In other words, they communicate for the sake of communication and to remind each other that they care. Social arrangements make up only 11% of emails and 12% of texts, suggesting such arrangements are made in other ways-presumably in person. Before digital technology, such a large amount of relationship work would require being on the phone or physically present with one's partner. Now, couples can engage in this work more often and while doing other things like sitting in class, hanging out with friends, or working-all completely isolated from one another [89].
Substantive techniques. Similar to other intimate ties, participants rarely qualify their romantic partnerships. Though occasionally they call partners by nicknames or terms of endearment (11% of emails, 4% of texts), this is surprisingly uncommon. The most frequent information types in email are humor and sexual subjects (both 16.4%), whereas in text messages banter (17%) is most common-painting a lighthearted portrait of digital exchange. The most frequent emotion in emails is romantic expression (13.7%) and in texts, support (9%). Unsurprisingly, sexual subjects and romantic expression were more common among partners than all other relationship types.
Discursive techniques. Participants used few discursive techniques with their partners; only 9.6% of emails and 8% of texts contained any. However, emoticons are used more frequently between partners than any other relationship type. Because these communications are already so light-hearted and relation-oriented-and because partners are more comfortable expressing emotions explicitly-much of the need to rely on discursive (as opposed to substantive) devices may be eliminated.
Social functions. Over half of messages between partners confirm intimacy (65.8% of emails, 53% of texts). As evidenced by the distribution of purposes, participants use digital communication to stay close to their partners throughout the day. Through notes of affection and humor, they maintain the strength of their connection even while apart. These messages also increase solidarity (15.1%, 18%); by expressing agreement, partners affirm elements of a shared social consciousness. Examples.
ugh it's pouring! guess the gloomy weather means we have to snuggle all afternoon. . . bummerrrrr: D call me when you're out (M text to romantic) Hellooo pup. Sorry we didn't get a chance to talk this morning. Currently am in the car on my way to polo. Miss you very much and can't wait to talk to you tonight! Have a great day. Hugs and kisses also: http://5napkinburger.com/ coming to boston (!!!) (F email to romantic)

Discussion
In this section, we consider the implications of our results in terms of the nature, relational patterning, and societal impact of text-based digital interaction. Some of the points we make have received little or no prior attention; others support, confirm, or expand upon findings from prior research.

Dramaturgy in the digital age
In the voluminous literature on CMC, many others have commented on the distinct nature of digital (as opposed to face-to-face) exchange. One challenge of this research is that "digital" takes so many forms, creating obstacles to triangulation. Particularly notable in this regard are social media [50,51], where the richness of interaction possibilities on any one platform-and tremendous variation across them-make precise characterizations of any one "medium" all but impossible. Further complicating matters is that many such sites no longer exist (or have changed dramatically over their lifespan), making it difficult to pinpoint how each functioned, when. In contrast, our focus on email and SMS helps clarify which findings may be unique to certain platforms or potentially innate to text-based exchange. Like many scholars before us, we find dramaturgy useful as an organizing lens. An elevated personal front. According to Goffman [36], front is comprised of a performer's 1) setting and 2) personal front, or the self-image she projects. Because there is no physical setting in digital exchange, the role of the personal front is substantially elevated. At the same time, many of its features-including body language, facial expressions, and current appearance-are unknown. Thus, it is at once interactants' only source of information and depleted of many of its detail-providing capabilities [28].
The consequence of this shift is that manner is reduced to explicit expressions. Face-to-face manner is frequently unintentional: Someone is excited, so her eyes light up. While such mannerisms are often strategic-like staring at a blackboard while thinking about the weekendthey are always deliberate in digital exchange. For instance, a digital interactant might write, "Yayayayayayayay OMG OMG I LOVE YOUUUU" (DT.90). Here, traditionally tacit information (great enthusiasm and friendliness) is shared through intentional decisions (capitalization and extra letters). Impressions "given off" are thus in fact impressions "given" because all parts of a presentation are consciously chosen [26,34].
Two implications are notable. First, because digital personal fronts are so important-and presumably deliberate-instances when rules about them are violated may be particularly jarring. Nervous laughter can't be helped, but explicitly typed laughter ("hahaha") at an inappropriate time could seem offensive. Similarly, just as it would be strange to yell a private matter, communicating sadness in all capital letters would (almost humorously) undermine the performance of sorrow. Successful face-to-face performances require that appearance, manner, and setting are all congruent. It is equally, if not more essential to digitally communicate with consonant substantive and discursive techniques-reflected in the great care with which messages (especially to authorities) were constructed and even "proofread" (Interview 2).
Second, however, some people may prefer this explicit articulation of meaning under challenging or uncomfortable circumstances like corrective face-work [112]. On one hand, apologies (or excuses) can be carefully crafted to contain the right tone and detail (e.g. IP.15 and the illustrative message provided at the end of the section on authority figures). On the other hand, one can avoid the emotions face-to-face interaction may evoke-such as the embarrassment of looking someone in the eye following a role violation-and the challenges of both verbally and non-verbally performing remorse. As one respondent summarized, "It's easier to say how you're feeling when the person is not sitting right in front of you" (Interview 8). The physical distance and temporal flexibility of textual interaction may be advantageous in this case, insofar as individuals are more likely to apologize. Similarly, another participant expressed a preference for receiving unexpected or potentially challenging communications this way: "Say I hurt someone's feelings, I would rather them text me so I can figure out what I did wrong and be prepared to apologize, rather than if someone is visibly upset and I'm caught off guard" (Interview 10).
Greater temporal complexity. Because email and SMS are asynchronous, time is important beyond the nuances of individual mannerisms [58]. While face-to-face exchanges have their own temporal rhythms [113], pauses are generally short and some response can be assumed. In contrast, digital response times vary widely-from moments to minutes to months-and can be motivated by a wide variety of intentions or circumstances-from being in class with your phone on "silent" to intentionally ignoring a message. Consequently, while substantive and discursive decisions are viewed as deliberate, the issue of whether or how time is meaningful is a matter of constant ambiguity.
Exemplifying this is when a response is absent or substantially delayed. Although short response times are typical among intimate ties, it is universally understood that people may take a while to respond. (There is also a general leniency awarded to authority figures. One participant accounted for this with the power dynamic between professors and students, bemoaning, "You are beholden to them so you can't complain"; Interview 14.) However, during rapid exchanges-of texts or emails-a sudden change of pace may be interpreted to reflect frustration or anger. As one participant elaborated, "I mean the person on the other end clearly asked a question or said something to you. And it's like, 'I clearly read this because I always have my phone and I'm not responding.' It's pretty obvious they're pissed" (Interview 2). In this way, such delays are not only symptomatic of becoming angry and not wanting to talk any longer-they are established non-verbal mechanisms of communicating these emotions.
And yet, long response times can mean many other things. They can be attempts to seem cool and easy-going, particularly when people are flirting or just becoming friends: Though the recipient may anxiously await each message, a delayed response attempts to conceal this fact-giving a relaxed or indifferent impression (Interview 4). They can also be unintentional, such as when an interactant is interrupted by someone in person. In this sense, it is striking how accurately participants interpret delays based on past information and situational cues; physical separation hardly prevents keen attunement to moods and communication patterns among intimates.
An expanded backstage region. With the transcendence of physical space comes a fundamental redefinition of performance regions. Because a digital message can be composed from anywhere (and typically out of sight of the audience), the entire performance is created and remains backstage (in the sense of a physical space where one can engage in discordant behavior) until the moment the "send" button is tapped. While this enables a variety of situations impossible offline-such as instantly jumping between roles in different messages-we focus on two particularly important implications: First, actors have increased resources at their disposal-including not only time [114], but anything available in their current setting that could aid the projection of front, in ways faceto-face presence could never allow. Several subjects confessed to consulting dictionaries, newspapers, and websites when writing messages. The potential tools for digital presentation of self are virtually limitless and extend far beyond one's own knowledge and front-repertoireincluding the knowledge and repertoires of other people.
Second, then, the redefinition of regional boundaries creates new possibilities for team collaboration. Every participant in our study admitted to sharing emails or texts with friends and both giving and receiving advice on how to respond. One message can have multiple authors (and multiple audiences-see below). Consequently, teams can collaborate not only to preserve a group's image, but also the image of a single person. Another new possibility is discreet communication in the presence of others. Face-to-face collusion is often tacit, as team members work off one another but can only openly discuss their line beforehand or afterwards. In digital interaction, actors can discuss and defy their performance in the very act of performing it, such as the erotic text sent "at work" in ST. 59.
Portable, permanent performances. While the backstage region expands, the front stage region shrinks and expands in digital interaction. On one hand, the recipients of emails and text messages are deliberately chosen (whereas in person, it is not always possible to isolate one's intended audience). In this way, digital interaction is highly privatized. On the other hand, offline performances are limited to the space and time in which they occur. Digital performances, in contrast, are effectively fossilized: Once they have been "performed" (i.e. sent), they exist independently of the performer-who loses all control over (and awareness of) potential audience members. Further, messages can theoretically exist forever and are easily disseminated [112]. In other words, unlike face-to-face performances-which tend to be stationary, temporary, and limited in scale by space constraints-digital performances are portable, permanent, and potentially public (Fig 1).
This results in a new and more self-aware type of performance. While performers present a front that is appropriate for the current message, they often do so in a manner that would not threaten their face in other relationships should the exchange be shared. In ST.71, for example, the participant's mockery of his cousin is implicit-he does not state that he is sending the link at his cousin's expense. In ST.73, though the interactants talk about a girl, they are careful never to state her name should someone outside their circle happen to see the communication. Explained one subject, "I try not to say people's names if we're talking about them [in text messages] because I'm paranoid I'm going to lose my phone" (Interview 12). Other interviewees reported experiences of disaster when past digital conversations were made public (Interviews 6, 7, 11). Whereas embarrassment in the present is a policing factor in face-to-face interaction, the potential for future embarrassment restrains digital exchange.
We contrast these dynamics with "context collapse" [45], where differentiation in self-presentation is impossible because multiple audiences are flattened into one. Davis and Jurgenson's [115] concept of "context collision"-the unintentional merging of audiences-may be more relevant. Meanwhile, these findings are otherwise similar to boyd's [116] characterization of social media (especially her concepts of "persistence," "scalability," and "invisible audiences"), underscoring the importance of differentiating between general properties of "bit-based content" (p. 46) and the specific manifestation of these traits in "networked publics" [117].

The relational patterning of digital communication
Emails and text messages are used by people of all ages and backgrounds. A basic feature of social life is that the substance of communication is tailored to the audience at hand. However, how the properties of digital interaction vary by relationship type has received little systematic attention [89].
While our data also showed important variation by medium, we were surprised to find it was not more pronounced. Further, such differences stemmed primarily from technological constraints: Compared to emails, texts convey less information due to character limits, lack space for conventional organization, and are harder to compose; consequently, they are more likely to be used when information is simple and users desire a fast response. Importantly, however-and plausibly due to these traits-SMS is also viewed as more casual [118]. When asked why they seldom texted authority figures, participants did not cite inability (though presumably few had access to the cell numbers of their professors or TAs). Instead, they indicated it would be "too informal" (Interview 1) or just felt "wrong" (Interview 8).
This qualitative impression (that the medium is less determinative than the relationship) was supported by a supplementary quantitative analysis performed using UCINET [119]. Specifically, we considered the summary statistics presented in Table 5-a matrix of 47 codes (rows) by 10 relationship types (columns)-as the raw data for a cluster analysis. First, we calculated the distance between each pair of relationship types as the Euclidean distance between the relevant two vectors of code frequencies (measured as percentages). This generated a 10-by-10 matrix of dissimilarity scores. (Alternative measures of distance produced similar results.) Second, we ran a hierarchical clustering algorithm on these dissimilarity scores based on a weighted average approach, i.e. where the distance between clusters is the average dissimilarity value weighted by cluster size. The dendrogram from this analysis (Fig 2) shows that  Table 5. The dissimilarity among types was measured using Euclidean distance and the clustering solution was produced using a weighted average approach. Analysis was performed using UCINET [119].
https://doi.org/10.1371/journal.pone.0273726.g002 based on the distributions of codes, communications sent to the same category of recipient are consistently more similar than communications sent using the same medium.
To our knowledge, this similarity between texts and emails-conditional on the nature of the relationship-has not before been identified, leading to a potential over-emphasis in the literature on the distinctiveness of each. In fact, among all 47 major and minor codes we identified, only two ("introduction" and "reference to past encounter") were unique to a certain medium (email).
While communications in each relationship category were characterized by a unique constellation of traits, the primary schism we observed was based on the intimacy of the relationship-where messages to family, friends, and partners (i.e. more intimate ties) were generally similar, as were messages to peers and authority figures (i.e. less intimate ties). Communications with authorities show respect through deference, gratitude, and formal address and primarily function to heighten solidarity and maintain face. Digital interactions with peers are most comparable, particularly with respect to their functions and formal style. In contrast, communications among intimates are lighthearted, loving, and supportive; they perform relationship work and confirm intimacy; and discursive techniques are common.
Within these two groups-low intimacy and high intimacy-the element most predictive of variation in message properties is age. Specifically, when communicating with peers as opposed to authority figures, or friends and partners as opposed to parents (the majority of recipients in the "family" category), subjects are engaging fellow members of a generation characterized by strong digital fluency-people who share a nuanced technological and cultural understanding of "how" to use modern devices. Accordingly, messages to bosses the same age as respondents looked very similar to messages to peers.
Finally, though digital communication transcends space, the physical distance between interactants still impacts the form messages take [120]-though our data suggest it is the weakest predictor of the three. In contrast, Johnson et al. [73] found little impact of distance on college students' relationship maintenance strategies; Boneva et al. [17] found men and women have different patterns of long-distance communication; and Mok et al. [96] found frequency of contact is generally insensitive to distance, except for transoceanic relationships. (All three studies examine email, not SMS.) While all five of our relationship types can be generally classified by intimacy and age, the distinction between proximate and distant relations arose inductively. Of course, proximity is likely correlated with other variables (such as how frequently the sender and recipient tend to see one another); accordingly, some message features (e.g. practical arrangement) are less relevant for distant ties.
Moving from the most foretelling property (intimacy) to the least (physical distance), it also appears that a "distant" value on any one relational component limits the predictive power of those that follow. For example, if the sender and recipient share a casual relationship, age has limited impact and the influence of physical distance is negligible. If the relationship is intimate, however, age increases in importance: A text message to a parent may differ markedly from one to a sibling. Lastly, proximity matters most for intimate, same-age ties, and has less impact on messages to older recipients (Fig 3).

Digital exchanges and modern society
Lastly, we consider potential byproducts of the widespread use of SMS and email on select features of everyday interaction and on the structure of social networks.
Sociability, stability, and solidarity. Although digital technology has transformed social life in many ways, we emphasize three consequences implied by our data. First, the accessibility of digital communication undeniably impacts the frequency of certain forms of interaction -making some, like sociability [121], easier to perform. Today, if someone is bored (e.g. during "idle times" like riding public transit [60,122]), she can text a friend to pass the time (not unlike listening to a Walkman or playing a Game Boy not long ago-functions also absorbed by contemporary smartphones). Digital communication facilitates communication for communication's sake; it is out of this utilization of texting, for instance, that banter like ST.66 and ST.74 are born and (within our "intended purpose" family of codes) non-practical information is exchanged.
Second, however, this ease of access has its costs. A central example is micro-coordination. As a rule, it is inessential-it is a tool of politeness and convenience borne of technological capacity. But because cointeractants can always be reached, plans are always subject to lastminute change; the availability of digital communication thus serves to both confirm and undermine plans until the very moment of their execution, decreasing the stability of arrangements [123].
Third, the theme throughout our data of shared sentiment in social groups naturally invites reference to mechanical solidarity [124,125]. Explicitly captured by our "increase solidarity" code, digital channels facilitate information diffusion and shared knowledge; exchanging links, pictures, and jokes elicits common emotional reactions, magnifying feelings of closeness; and shared distaste towards others (e.g. gossip or derision) further affirms the consciousness of the in-group. That email and SMS are more often used to emphasize similarity rather than difference could also reflect the constraints of text-based exchange (where nuance is inconvenient, Fig 3. The predictive importance of intimacy, age, and proximity. The influence of a relationship feature on the structure of a communication decreases the further to the right and the more transparent that feature appears in the above diagram. Thus, opaque features are more important and transparent features less so; and intimacy is more important than age, which is more important than proximity. https://doi.org/10.1371/journal.pone.0273726.g003 emotional monitoring is difficult, and odds of misinterpretation are high-unfortunate staples of online debate that close ties might prefer to avoid).
Broader and deeper social networks. Digital exchange has low costs and high returns. It does not take much time to craft a short message, yet such basic acts of communication (from well wishes to romantic expressions, check-ins to "for yous. . .") enrich and revitalize social ties [89,97]. This important "social function" was also explicitly identified by our coding. On one hand, the widespread use of SMS and email make it incredibly easy to communicate with people we are unwilling or unable to contact by traditional means-including introductions (e.g. ST.28, ST.42, and Interviews 2 and 12). On the other hand, young adults today are saturated with an endless flow of messages from family, romantic partners, and friends, even when they are physically distant-maintaining high levels of intimacy that might have faded in a previous age [8,22]. Consequently, digital interaction facilitates both the maintenance of weak ties outside our normal social circles as well as the frequent activation of strong ties near or far [89], effectively "broadening" and "deepening" our social networks (Fig 4).
These changes entail benefits and challenges alike. A large body of literature documents the advantages of CMC, including increased authenticity and self-disclosure [26,126,127]. Digital interaction can also facilitate other forms of exchange: Friends can be invited to join someone at a bar (IP.21); plans can be made to catch up by phone (IP.24) or other "logistics" can be arranged (IP.1-6); and because individuals do not have to reestablish lines when they have stayed in touch digitally, face-to-face interactions are smoother. Indeed, the more time that has passed between offline encounters, the larger the possible gulf between performancesand the greater the utility of digital exchange for bridging this fracture.
But there are limits to network engagement-even digitally [128]. In a sense, young adults are so frequently in touch because there is no excuse not to be. However, such communication detracts from other experiences-including private activities, offline interaction, and valuable (but increasingly scarce) moments without external stimulation [129]. Many relationships now subsist on validation in a way they could not, previously. Rather than unspoken and enduring bonds, such ties must be constantly tended to avoid disintegration-a delicate (and potentially draining) endeavor.

Conclusion
Digital communication is now so accessible and widespread that any portrait of modern social life seems incomplete without it. This is particularly true for young adults: 48% of 18-to 29-year-olds say they go online "almost constantly" [130]. On one hand, the popularization of the smartphone-which makes email immediately accessible-and especially the iPhonewhich enables "text messages" (via iMessage) to surpass 160 characters-has made emails and texts increasingly indistinguishable, potentially blurring the differences we observed. On the other hand, these mediums are now richer than ever before (including not just pictures and animations, but meta-communication such as "likes" and "laughter"); additional features (such as read receipts and "typing. . ." notifications) further complicate the temporal dynamics described above; and the plethora of available platforms (from Instagram to Facebook, What-sApp to Zoom) mean that emails and text messages are but two options on an ever-diversifying communicative menu. For all of these reasons, it is that much more important to have an analytic framework for comprehension and an empirical baseline for comparison.
To what extent did we achieve these ends? Given the peculiarities of our research design, it seems appropriate in closing to return to the core issues of generalizability our data raise: namely, the extent to which self-selected, text-based messages furnished by a convenience sample of college students in 2010 are relevant to a broader population of people and the many digital interaction forms available today.
First, we believe our research would not be possible with a random sample of participants or messages; indeed, that anyone was willing to provide such an intimate look at their private lives was fortuitous (and our response rate was still only 65% among initial contacts). Though supplied by 15 respondents, our vast dataset of messages generated a nuanced and transposable theoretical framework that meets evaluative standards for our method [100,131]. Comparisons by medium and relationship type established vibrant empirical patterns, even if precise code distributions are ungeneralizable. Future research could identify additional codes our research design precluded; examine other kinds of relationships beyond (or within) the five examined here (for instance, colleagues, adversaries, or partners of varying duration or commitment); or even use our results to seed automated analyses of large corpora without examining individual messages, simultaneously protecting author privacy and enhancing scale [132].
Second, our methods have enabled us to show that some results produced by studies of specific digital platforms (often social media) are in fact generic, helping bridge the gap between broad, classic theories of computer-mediated interaction [26,27] and empirical work. Insofar as all communication contains content, our framework should be broadly useful-although non-textual and/or synchronous forms may suggest entirely new analytic properties. As digital interaction continues to evolve-replicating, if not surpassing, more features of face-to-face exchange-it will be increasingly valuable to identify precisely which interactional factors are responsible for what substantive consequences. At the outset of this paper, we suggested some basic dimensions of variability, including communicative affordances (e.g. textual vs. audiovisual, synchronous vs. asynchronous), social structure (e.g. dyadic vs. group, familiar vs. anonymous), and technical interface (in terms of ease, accessibility, and length). While other dimensions are possible-including how each manifests in distinct cultural and historical settings-we suggest that any such typology of digital forms might usefully begin with the elementary.
Ironically, given the explosion of "big data" research examining large social systems in microscopic detail [133], digital scholarship has yet to fulfill one of its most enticing possibilities: providing new insight into the content of everyday exchanges [134]. Despite its limitations, we hope our work will be perceived as an important and timely step in this direction.