Illocutionary acts and functional orientation of SMS texting in SMS social networks

Susana M. Sotillo
Montclair State University

Abstract

1. Introduction

Mobile telephony and computer-mediated communication have been studied from multiple perspectives in a variety of disciplines including behavioral psychology, communication studies, discourse analysis, sociolinguistics, social network analysis, and sociology. With respect to the impact of mobile devices on language and communication, Baron (2008) has shown how electronically-mediated technologies are changing the way we communicate and relate to each other. More extensive findings relating to relational and linguistic features, including the maxims that speed and brevity, paralinguistic restitution, and phonological approximation characterize SMS texting, are presented by Thurlow & Poff (2011). In addition, Tagg’s (2009) doctoral research shows that while not all text messages are brief, texters display creativity and performativity in their texting practices.

There are several studies of short message service (SMS) texting practices from a social network analysis perspective that have measured adolescent use of socially interactive technologies (SITs) and their relationship to offline social networks (Bryant, Sanders-Jackson, & Smallwood 2006). Their results showed that socially isolated adolescents did not use SITs, but those who favored their use did not necessarily increase their offline ties. Other studies of teen and young adult texting and instant messaging have also addressed social functions and linguistic and metalinguistic issues. Contrary to the exaggerated claims of journalists and social pundits, a positive association between texting and literacy practices, including better performance on standardized measures of English proficiency, was found by Plester, Wood, & Bell (2008) and Plester, Wood, & Joshi (2009). Texting was shown not to interfere with the acquisition of the standard code (Tagliamonte & Denis 2008).

Scholarly research as well as the popular press point to the importance of text messaging or digital communication, which is widely used around the world by children, adolescents, and adults. Anyone, anywhere in the world has the ability to send a message via the SMS of mobile phones, smartphones, and other techno-cultural artifacts. Opencode-Mobile Network Systems (February 2012) estimates that the total number of text messages sent by individuals/consumers to mobile phones worldwide is approaching 10 billion SMSs per month. This technology, according to ABI Research, continues to be the preferred mode of communication among younger groups, with the most popular types of communication being SMS, MMS, mobile email and instant messaging. Their use will vary according to age, purpose, and affordability (Tsirulnik 2010). In addition, SMS today is widely used in the corporate world for instant notification of important events.

In the United States, not all teenagers, young adults, or individuals with limited income are in a position to afford the mandated data plans when purchasing communication devices. However, these groups, including mature adults, are in a position to use SMS texting for maintaining social relationships, obtaining information, scheduling appointments, asking/answering questions, criticizing, or producing the textual equivalent of “illocutionary” acts (i.e., meaningful utterances/messages with a certain conventional force) or texter’s intention in sending a text message. Texters also produce what are in essence “perlocutionary” acts that compel the message recipient to react to the message received. For example, a recent study by Nastri, Peña, & Hancock (2006) used speech act analysis to examine the nature of “away messages,” a form of digital communication. Their results showed that assertives and expressives were more numerous than commissives, and that directives were rarely used because away messages are constructed for informational and entertainment purposes, confirming previous findings by Baron et al. (2005). In the case of SMS texting, studies have shown that people send texts to those who form part of their personal or professional SMS social networks for a variety of reasons (Betti 2008; Ling & Donner 2009; Lloyd & Gillard 2010; Sotillo 2010; Tagg 2009; Thurlow and Brown 2003; Zarantonello 2001). These include constructing identities, enhancing linguistic creativity, strengthening personal and professional relationships, exchanging information, and maintaining SMS social networks.

This study builds on previous research by Nastri, Peña, & Hancock (2006) and Twitchell & Nunamaker (2004) who used SAT as a model for ­analyzing language use in computer-mediated communication because it allows researchers to explain the intended meaning of speech acts. However, few studies have used SAT in order to analyze digital communication in SMS texting and this is the purpose of this investigation. The study also seeks to expand on Thurlow and Brown’s (2003) findings concerning the sociability or relational nature of SMS text messages. SAT and Thurlow and Brown’s framework are useful for analyzing SMS texting data since these two approaches complement each other. By using SAT it is possible to explain the intended meaning of text messages identified and coded as different types of illocutionary acts. Similarly, an adaptation of Thurlow and Brown’s (2003) framework is helpful in uncovering the communicative intent or functional orientation of these text messages. By combining these two approaches the aim of this exploratory study is to explain how illocutionary acts are instantiated in text messages, and what their intended function or purpose is. Two general research questions are addressed: 1. What types of illocutionary acts are found in the texting practices of individuals who form part of six SMS social networks? 2. What is the communicative intent or functional orientation of illocutionary acts evident in these SMS texting exchanges?

2. The corpus

2.1 Population

This study is based on a subsample of a corpus of 5,809 sent and received SMS text messages donated by six individuals, three males and three females, who owned six SMS social networks. Data collection was approved by the researcher’s Institutional Review Board. Each SMS social network centered around an individual and his/her texting partners, which included family, friends, relatives, neighbors, and co-workers. The six who contributed their text messages were teenagers, colleagues, neighbors, friends, and some family members of the researcher. The individuals identified as the owners of the six SMS social networks included two female teenagers, ages 16 and 19, respectively, and a 17 year-old male. The adult SMS social network owners included a 57 year-old female educator, a 38 year-old IT specialist, and a 62-year old male professor. SMS texting social networks in this study are defined following Fuhse (2009) as “configurations of social relationships interwoven with meaning” (p. 51), which are comprised of sociocultural structures. In ego-centered networks, the structure of relationships can be examined by looking at the types of transactions that take place between individuals (Fuhse 2009: 52). Thus, an SMS social network is defined as a complex system of interpersonal texting relationships and expectations between an actor or texter (SMS social network owner) and various individuals who receive and respond to the texter’s messages. More than 59 individuals from different social backgrounds and ages formed part of the six SMS social networks. In the subsample analyzed, the average number of participants in the three teenagers’ SMS social networks was 19, whereas it was 23 in the adults’ SMS social networks.

2.2 Data collection procedures

Sent and received text messages collected for this study were downloaded by contributors using software developed for smartphones such as Treos and Blackberries, among other brands. The subsample consisted of 1,271 sent and received text messages that were randomly selected from the beginning, middle, and end of the SMS text message data sets contributed by each of the six owners of the SMS social networks. There were 417 sent and received text messages in the SMS social networks of the teenagers, 348 from owners and members of two of the adult SMS social networks, and 506 from information technology (IT) specialists, who formed part of the SMS social network of the third adult contributor.

3. Theoretical framework of analysis

Though most studies motivated by Speech Act Theory (SAT) have focused on the utterance as the unit of analysis, digital technologies require new perspectives or approaches to the study of language, which in the past have often privileged the spoken language or literary language. SAT as a method for ­analyzing SMS texting data was chosen because it is a practical taxonomy that has been used in studies of natural language processing and computer-mediated communication (Nastri, Peña, & Hancock 2006; Twitchell & Nunamaker 2004). Furthermore, as mentioned in the Introduction, SAT allows researchers to explain the intended meaning of text messages identified and coded as illocutionary acts, whereas approaches such as Discourse Analysis and Conversation Analysis generally focus on the co-construction and negotiation of meaning in face-to-face interaction or recorded telephone conversations.

Few studies have used SAT or related taxonomies to analyze digital communication in Instant Messaging or SMS texting (Nastri, Peña, & Hancock 2006). However, there is still a need to study the pragmatics of text messaging practices. This study seeks to add to recent findings concerning the nature of SMS texting by using SAT analysis and an adaptation of Thurlow and Brown’s (2003) communicative intent-functional orientation framework. In SAT, functional units of communication have locutionary or propositional meaning (i.e., the literal meaning of what is said), illocutionary meaning or the intended meaning of what is said, and perlocutionary force or the effect of what is said on a hearer or message recipient. In this study, the focus will be on illocutionary meaning because the data consist of text-based digital exchanges. Also, the functional orientation of these text messages will be investigated.

3.1 Brief overview of SAT

Communication and language scholars often draw on Speech Act Theory (SAT) in an attempt to explain how speakers use language in order to accomplish intended goals or actions and how hearers or recipients of an utterance or message infer the intended meaning from what is said. Austin (1962) initially presented five general classes of illocutionary acts or conventional force of utterances (verdictives, exercitives, commissives, behabitives, and expositives). His concepts and taxonomy were extended by Searle (1965, 1969, 1975), who proposed constitutive rules and broadened Austin’s notion of uptake by asserting that the performance of an illocutionary act involves the securing of uptake (Halion 1989).

In his comprehensive review of the history of SAT, Smith (1990) traces back the development of speech act theory by analyzing the contributions of various philosophers, beginning with Thomas Reid’s rejection of the Aristotelian view of language toward the end of the 19^th century. He expounds on the insightful contributions of various members of the Munich school of phenomenologists (1990: 49) and demonstrates the closeness of Reinach’s thinking to Austin’s and Searle’s. Searle, expanding on Austin’s work, introduced a distinction between literal word or sentence meaning and speaker’s utterance meaning. For Searle, the production of a word or sentence in the performance of a speech act constitutes the basic unit of linguistic communication (Searle 1965: 136).

SAT is not without problems because of flawed conceptualizations of language that regard the sentence as the unit of analysis for speech acts (e.g. Austin 1962; Searle 1969, 1975, 1979) and the claim made by some scholars that there are universal pragmatic principles (e.g., Austin 1962; Habermas 1976; Brown & Levinson 1978). Others argue that speech acts are culture-bound notions that vary across cultures and languages (Allwood 1977; Rosaldo 1982; Wierzbicka 1985). This investigator finds that this criticism is well founded in that studies of speech acts in other cultures (Rosaldo 1982) can only be interpreted within the context of prevailing “situational and cultural constraints on forms of language use (Rosaldo 1982: 203). Additionally, Searle’s categories for analyzing speech acts have been criticized by scholars such as Burkhardt (1990) and Clark (1996) because of his assumption that each speech act can be neatly placed in one category, ignoring the multiple functions of language. Understandably, this is not always possible, especially with respect to SMS text messages since the intended meaning of illocutionary acts can be ambiguous to an outsider examining text messages from a particular SMS social network of which he/she is not a member. Thus, inferring meaning and pragmatic functions from SMS text messages is a challenging undertaking.

4. Analysis of the data

4.1 Classifying illocutionary acts

Searle’s (1969, 1979) basic taxonomy of five types of illocutionary acts is used in this study in order to identify, classify, code, and subsequently tag speech acts for processing using MonoConc Pro 2.2 (Barlow 2002), a software tool. The text message as an illocutionary act is defined as a propositional unit, which sometimes includes punctuation. Linguistic routines such as greetings and leave-takings, expressives, phatic, and conative interjections, and onomatopoeic speech act verbs, a subset of interjections, are also categorized as speech acts (Tsai & Huang 2003). Thus, notwithstanding their nonsentence status, greetings and leave-takings that consist of more than one word (hi how are you), emoticons and some common acronyms, initialisms, and abbreviations (e.g., LOL, Pls, idk, flmao, kk) were coded depending on the context of the exchange as either part of expressives, assertives, or some other category. Digital communication researchers have shown that these linguistic forms are considered part of a global SMS language (e.g., Ling 2007; Thurlow & Poff 2011; Deumert & Masinyana 2008). Table 1 illustrates how Searle’s five types of illocutionary acts (assertives or representatives, directives, commissives, expressives, and declarations/declaratives) were used to code and analyze the SMS texting data.

4.2 Tagging and analysis of illocutionary acts

Downloaded text messages were manually coded and tagged according to the categories described in Table 1. Initially, we sought to identify performative verbs in the text messaging data based on Vanderveken & MacQueen’s (1990) analysis of performative verbs by generating lists of verbs using MonoConc Pro2.2. However, this was not a productive endeavor as few of the performative verbs identified by Vanderveken & MacQueen were found in the subsample (e.g., agree, invite, promise, tell). The illocutionary act or intended meaning of speech acts in text messages had to be inferred from context. Text messages coded and tagged as illocutionary acts represented propositional units carrying the intended message and were often but not always separated into different units of communication by punctuation: <ASS>Just hd Palm Dsktop pblm</ASS> <ASS>& gt gd hlp fm lady tech-in the Philippines!</ASS> <DIR>Ask her about the software to save cell text messages.</DIR>. Each text message was examined carefully since some contained two or more propositions. For example, “hates Kenny, etc in Belleville” “husband was R's aide.” “I guess he's back on staff.” Though these constitute three different illocutionary acts, they formed part of one long text message. In general, it was easier to code mutually exclusive categories when unambiguous examples could be found (e.g., directives vs. expressives): <DIR>Please do not go down this road right now.</DIR> versus <EXP>thanx bro</EXP>.

Once text messages in this subsample were coded and tagged, lists were generated for further analysis. Since IRB guidelines required that the investigator remove all identifying information (e.g., cell phone numbers, email addresses, first and last names or other indentifying information), coding for illocutionary acts in this subsample was at times very difficult because adjacent and functionally related turns were not easy to identify. Nevertheless, sequences of related messages by different texters were found primarily in text messages donated by one the three teenagers and in those of most adults. (see untagged mini example)

Text messages representing propositional units ranged from six to 75 characters because only 160 characters are allowed per message. Thus, if a texter sent a very long message that included “if-then” clauses or several propositional units, these were coded separately as illocutionary acts and often appeared as two or three consecutive text messages. There were 10,840 words in the 1,271 text messages analyzed, which yielded an average of 8.53 words per text message in the corpus after removing punctuation and treating contracted words as one (e.g., “we'll,” “can't”). Also, an average of four intervening messages were found between a question or request sent by a texter and the recipient’s reply, except for the messages contributed by a 38-year old male. In his SMS data there were 21 intervening text messages between a request for information by a member of his own SMS social network and the answer he subsequently provided.

As with most high-inference categories, a certain degree of ambiguity in the definition of speech acts/illocutionary acts categories is present and this calls for insider cultural knowledge. This complicates the coding task as some illocutionary acts overlap and could be coded as belonging to two different categories. Also, critics of SAT point out that placing utterances, in this case, text messages, into mutually exclusive categories ignores the multiple functions of language use.

Two trained raters and the researcher initially coded all text messages in the subsample as belonging to one of five categories, but modified these to four because no examples of declarations were found in the data. Though there were disagreements with respect to the classification of illocutionary acts in some of the text messaging data, inter-rater reliability was statistically significant at p <.001) (Kendall’s Coefficient of Concordance=80). Illocutionary acts that were difficult to assign to either one or another category were discussed in face-to-face meetings until a resolution was reached.

4.3 Analysis of communicative intent-functional orientation

Communicative intention as defined by Grice (1969) was understood as the intended effect on a hearer or audience, and awareness of this intention that someone’s utterance had but this notion was criticized by Strawson (1964), Searle (1969), and others working within a formal mentalistic approach. For our purposes we define communicative intention as the intended function or purpose texters sought to accomplish through specific illocutionary acts (e.g., inform, order, complain, criticize, request personal favors, commiserate, or change behavior through perlocutionary acts). Sequences of text messages classified as assertives, expressives, directives, and commissives in this corpus were examined by the researcher and two graduate students and coded according to the specific communicative intent-functional orientation we inferred. After examining 57 pages of text messages or 1,391 lines of text, we came up with the following coding scheme modeled after Thurlow and Brown’s (2003) communicative intent-functional orientation framework: Information Sharing, Requests, Scheduling and Planning, Friendship and Intimacy, and Social Network Maintenance. Those that were ambiguous were not coded because it was impossible to reach agreement. These are described below.

Information sharing: These were classified as illocutionary acts instantiated in text messages that focus on sharing specific information in response to a texter’s request or based on statements that need further elaboration (T stands for texter and numbers indicate different participants):

Requests: These illocutionary acts dealt with requests for personal favors (indirect requests) and direct orders or commands from family members, social network friends, or acquaintances from work that the message recipient do or not do something:

Scheduling & planning: These consist primarily of illocutionary acts that texters use to coordinate meetings at work, plan lunch or dinner dates, arrange shopping trips or pick-up times, and invite friends and acquaintances to local functions:

Friendships and intimacy: These are illocutionary acts focused on emotionally loaded issues discussed with intimates, and subsumes text messages of a romantic nature with sexual overtones found primarily in the teenagers’ SMS social networks:

Social network maintenance: These consist primarily of text messages sent by the SMS social network owners for the purpose of keeping in touch with family, friends, and acquaintances or with relatives living in other cities or states.

Ambiguous: This category included illocutionary acts instantiated in text messages where the inferred function or purpose could not be easily categorized. For example, the following assertions could be classified as information sharing, but their intended communicative function appears to be didactic. It seems as if the political operative making these assertions is not just sharing information with others, but is also assigning blame for political losses and teaching a lesson: <ASS>In every campaign, lessons are learned. </ASS> <ASS>I'm sure B learned that he should have been holding fundraisers for the two years that he was telling anyone who would listen that he was running for freeholder.</ASS> <ASS>The other lesson is that nowadays there is no such thing as volunteerism in campaigns.</ASS> <ASS>People need an incentive if you expect them to provide quality help in an election. </ASS>

The following SMS exchanges also illustrate that it is possible to place these messages in more than one category:

However, 66 % of the communicative units examined in this subsample were successfully classified as to their intended functional orientation by all three raters.

5. Results

Research question 1 sought to investigate the types of illocutionary acts found in the SMS text messages of members who formed part of six different SMS social networks. Table 2 presents frequency counts for each of the four types of illocutionary acts that were identified and coded. Though Searle’s classification scheme consisted of five mutually exclusive categories, we only identified four since there were no declaratives/declarations or verdictives as these are speech acts that can only be uttered, performed, or written by individuals in positions of authority representing institutions, such as judges, police officers, or employers.

Figure 1 illustrates the percentage of each illocutionary act category out of a total of 2,020 illocutionary acts identified, coded, and tagged in 1,271 text messages. Assertives account for 39% of all illocutionary acts, followed by expressives (33%), directives (20%), and commissives (8%).

Figure 2 shows the distribution of illocutionary acts by owners and members of each of the six SMS social networks. Of the four categories of illocutionary acts, teenagers used assertives 51.64% of the time, expressives, 34.57%, followed by directives and commissives (10.38% and 3.41%, respectively). The distribution of illocutionary acts in the sent and received text messages contributed by a 62-year old SMS social network owner was as follows: 44.54% expressives, 25.21% assertives, 22.69% directives, and 7.56% commissives. One of the adult SMS social networks included a number of political operatives who sent SMS texts to the network owner and multiple recipients. This group showed the following distribution: 51.16% assertives, 21.19% expressives, 17.57% directives, and 10.08% commissives. Another adult SMS social network was made up of IT specialists whose distribution of illocutionary acts was as follows: 46.66% assertives, 22.95% directives, 22.32% expressives, and 0.08% commissives.