Bernicotal JOP2012
Bernicotal JOP2012
Bernicotal JOP2012
Bernicot, J., Volckaert-Legrier, O., Goumi, A. & Bert-Erboul, A. & (2012). Forms and functions
of SMS messages: A study of variations in a corpus written by adolescents, Journal of
Pragmatics, 44, 1701-1715. 10.1016/j.pragma.2012.07.009
This research was supported by the French National Research Agency project ANR-08-
COMM-011-01 entitled "Adolescents and Computer-Mediated Writing: Prerequisites, Uses,
and Learning".
[email protected]; [email protected]
Abstract
The purpose of this research was to gain insights into SMS communication among
French-speaking adolescents. We analyzed the effects of the writers' characteristics (age,
gender, and SMS-messaging experience) on message length (number of characters and
number of words), dialogue structure (with or without an opening and a closing), and
message function (informative vs. relational). The SMS messages were produced in a real-
world situation. We found differences across writers’ characteristics for all the dependant
variables. The commonly reported distinctions between girls and boys were mitigated.
Moreover, for dialogical structure, the messages differed from those found in traditional oral
and written interactions since 73% of them did not have the conventional opening-message-
closing format (the opening and/or the closing was missing). The results are discussed in
terms of the specific characteristics that define the SMS register, and potentially relevant
approaches to be taken in future research are addressed.
1. Introduction
Within only a few years, written communication via SMS1 has become an increasingly
frequent part of our daily lives. This readily accessible technology enables asynchronous
elite. It later became available to a larger number of individuals when compulsory education
was instituted in Western countries. In 1950, oral communication by telephone became the
emerged as a serious competitor. The use of SMS messaging as one of the various kinds of
forums or online help forums, and written communication on social networks has soared
since the year 2000 (e.g., Facebook, Twitter) (Crystal, 2008; Danet & Herring, 2007; Ellison,
Steinfield & Lampe, 2007; Fernandez & Yuldashev, 2011; Herring, 1996a; Puustinen, Bernicot
& Bert-Erboul, 2011 ; Puustinen, Volckaert-Legrier, Coquin, & Bernicot, 2009; Savas, 2011;
Valkenburg & Peter, 2009; Volckaert-Legrier, Bernicot, & Bert-Erboul, 2009). These new
modes of communication arouse both fear and fascination (Thurlow, 2006), as was the case
over previous centuries upon the invention of printing, the telegraph or the telephone
(Crystal, 2008). Their written forms, and particularly those of SMSes, diverge from those of
traditional forms of writing. However, these phenomena are not entirely new: for example,
1
SMS is the abbreviation for Short Message Service. Some synonymous terms are text-messaging and texting.
4
abbreviations or rebus writing have been used since the very origins of writing (Crystal,
1996; Vignaux, 2003). In 2001, Crystal's presentation of text-messaging took up two pages of
his book Language and the Internet, whereas he devoted an entire book to the topic in 2008.
From the scientific viewpoint, research on SMS messaging has provided new data for
answering a number of questions about language acquisition and language use. The novel
linguistic material found in SMSes is particularly useful in pragmatic studies, which attempt
to relate linguistic productions to the communicative situation in which they are produced
(Austin, 1962; Bernicot, 1994; Bernicot & Mahrokhian, 1989; Bernicot, Comeau & Feider,
1994; Bernicot, Laval & Chaminaud, 2007; Grice, 1975; Herring, 2004; Noveck & Sperber,
2004; Searle, 1969; Verschueren, 1999). The goal in the new studies is to determine the
own (Ravid & Tolchinski, 2002). In this vein, Crystal (2001) stressed the importance of the
concept of "language variety", which applies to oral and written language, legal and scientific
mediated writing, including SMSing. Variations between language registers are systematic,
predictable, and conventional. A university professor giving a lecture will probably not talk
like a sports commentator reporting a football game. Working specifically in the area of
from other language-focused disciplines (particularly linguistics). At the heart of CMDA is the
threads, archives, etc.). For aspects of language studied, it is important to answer the
5
following questions: how do they vary depending on the context, how are they learned, and
how do they change over time? A faceted classification scheme for CMDA (Herring, 2007),
based upon both medium and situation factors, enables the placement of SMSes with
concerned, using an SMS one can send 160-character messages from a mobile phone to
another computer or mobile phone located anywhere in the world, in a matter of seconds.
SMS messages are used in a nearly synchronous way like traditional oral communication,
while still being asynchronous like traditional writing. Among the situation factors, we
generally consider the characteristics of the SMS users (for example, their gender or their
experience communicating through a given medium) and the purpose—the social function,
the goal of the interaction or, in other words, what the individual participants hope to
Research in this area has dealt extensively with the most conspicuous feature of text
messages, i.e., their peculiar orthographic forms not found in traditional writing. Most of the
neographic forms inventoried so far are novel spellings resulting from puzzlegrams based on
homophonic letters and numbers, phonological reductions, symbols, and acronyms (Crystal,
2008; Frehner, 2008, in French: Anis, 2007; Panckhurst, 2009; Rivière & Licoppe, 2005).
Recent studies have also looked into the relationship between SMS-specific forms, and
spelling in traditional writing, during the learning period (Kemp, 2011; Wood, Jackson, Hart,
Plester, & Wilde, 2011) or in young adults (Drouin, 2011; Rosen, Chang, Erwin, Carrier &
Cheever, 2010).
6
The main aim of our study was to look at several variables other than spelling, which
might be just as important in characterizing SMSes written by adolescents: the length of the
SMS messages, their structure, and their social function. In the following, we will present a
literature review with regard to these three variables. We will then present the research
2. Literature review
Most studies on SMS messaging have looked mainly at their orthographic content. A
few studies have also analyzed measures such as SMS message length, structure, and
functions.
SMS messages contain at most 160 characters. How do adolescents use this space?
Does age have an effect? Are there differences between girls and boys? Does SMSing
The first studies were carried out in English, Norwegian, Swedish, German and
isiXhossa in adolescents and young adults between the ages of 13 and 27. The results varied
according to the method used. In small groups of participants who were known to the
researchers, message length varied between 65 and 133 characters and 14 and 22 words
(Deumert & Masinyana, 2008; Döring, Hellwig & Klimsa, 2005; Grinter & Eldrige, 2001; Hård
af Segerstad, 2005; Thurlow & Brown 2003). As part of a phone survey carried out with a
large number of participants, message length was shorter: 29 letters and 6 words (Ling,
7
2005a, 2005b). With regard to the 13- to 19-year-old age group, Ling (2005a, 2005b)
emphasized that the girls produced longer messages than the boys (30 and 20 letters,
respectively).
When describing the traditional rules of oral interaction between two people, Goffman
(1967) insisted upon the importance of greetings and leave-taking. These rituals are
described as obligations which allow individuals to keep face (a positive social value) in a
given situation. To define the interactive outline of written language, Herring (1996b) offered
“closing epistolary convention”. Openings and closings are addressed directly to the
interlocutor; they are discursive markers indicating how the speaker situates him or herself
with respect to the addressee. Do we find this same structure in SMS interactions: the
opening (greeting one another)—the message (saying what one has to say)—the closing
(taking leave)? As seen by Rettie (2009) this three-phase structure is pertinent for
interest.
In a study carried out by Ling (2005b), only 10% of the messages contained an opening
Spagnolli and Gamberini (2007) showed that, in young adults, 13% of messages had an
opening while 35% had a closing. A study focused on adolescents showed the same type of
8
sequence was only very rarely used by 14-year-old adolescents (even when taking into
Greek-language participants transferred the messages stored in their cell phones (and
Information about the participants’ SMS use was also collected. Overall, 30% of the
messages included a closing. The analysis showed that closings are more present in the
disengagement from the daily frame of contact between close friends, and as a mark of
the potentially face-threatening act of refusing a specific request). Furthermore, closings are
The available studies are scarce and not analytic enough to draw any definite
conclusions at this point. However, contrary to what is observed in traditional oral or written
interactions, there are SMS interactions without an opening and/or a closing. This could be a
should be noted that in electronic mail, even if they are not systematic, openings and closing
are much more frequent (50% to 100% of emails have an opening or a closing) (Bou-Franch,
What is the purpose of SMS messages? The question of a message’s purpose is linked
to that of its addressee. With Japanese and French participants, Rivière and Licoppe (2005)
9
showed that SMSes are mainly sent by close friends of all ages. The authors also stressed the
frequency of within-family use between parents and their children, and the more limited use
of texting in formal or professional situations. The majority of the exchanges studied had
emotional or affective content. In cases where the SMS's function was to keep a relationship
going, it did not replace the telephone call. A new mode of communication does not take the
place of another, but rather fulfills different social functions (Harper, 2005). The research of
Reid and Reid (2005) and de Ischii (2006) carried out in the United Kingdom and Japan,
To address this topic of what SMS messages are used to say (or do), we can consider
the point of view advocated in pragmatic theories, which consists of determining the writer's
intention (the illocutionary goal of the message) (e.g., Searle and Vanderveken's taxonomy of
speech acts, 1985). Prior studies have taken an approach consisting of defining categories of
social functions that appear useful for describing the practices of users.
Past research has proposed a variety of categories for analyzing the functions of SMS
messages. The only study that reported an inter-coder agreement rate (Faulkner & Culvin,
2005) obtained a very low figure (15%), which highlighted potential problems regarding the
relevance of their categories and definitions. However, all of the studies agree on one
revolving around friendship maintenance are highly frequent (Deumert & Masinyana 2008;
Faulkner & Culvin, 2005; Grinter & Eldridge, 2001; Ling, 2005b) and for some authors, are
particularly so among female adolescents (Döring et al., 2005; Kasesniemi, 2003). Thurlow
and Brown's (2003) proposal to classify the functions on an axis ranging from "more
information, practical arrangements and chain letters. The relational category includes social
collected messages, 31% fell into the first category and 61% fell into the second.
3. Research questions
All of the studies mentioned above (except for those of Spagnolli and Gamberini in
2007, and Spilioti in 2011, described above in section 2.2.) ran across the same
computer file format, which means that the participants have to copy a transient text. This
situation poses the general problem of corpus reliability (the original messages are not
number of writers. Furthermore, the collection method was direct and automatic, thereby
avoiding all the errors and bias introduced by the manual copying of messages. Responding
to an advertising campaign in the media calling them to “give your SMSes to help science”,
the participants sent messages stored on their cell phones to a free number which enabled
the researchers to receive the messages in their original form and to organize them by
participant and by date. Within this new framework, our study attempts to respond to the
1. What are the characteristics of text messages with regard to length (an index
2. What are the structure and social functions of text messages (two indexes
linked to the situation factors of SMSes)? In terms of the structure of the message, we were
particularly interested in the linguistic markers indicating how the writer situates him or
herself with respect to the addressee: the presence or absence of a salutation as a message
opening and the presence or absence of leave-taking as a message closing. For the social
function, we focused on the informative or relational nature of the message. We used the
principle of classification from Thurlow & Brown (2003) and adapted it to our data2.
3. How do SMSes vary (in length, structure and function) according to the
message writer’s sociolinguistic characteristics such as age, gender and SMS experience (in
terms of how long and how frequently the person has been texting3)?
(ages 15 to 18), a period for which the little developmental information available is imprecise
or contradictory, despite the fact that this is the age range during which text-messaging is
used the most (Höflich & Gebhardt, 2005). SMS messages are both inexpensive and more
discreet than phone calls. It is also important to consider that the conventional written-
language register develops between ages 9 and 17 (Berman, 2005), and all aspects of written
language use evolve during this period (Jisa, Reilly, Verhoeven, Baruch, & Rosado, 2002;
2
The details of our coding can be found in section 4 (4.3.3. Function indexes).
3
Section 4 (4.1. Participants; 4.2. Procedure) and Appendix A provide a more detailed definition.
12
We also took the gender of the participants into consideration. Like the original studies
on oral language from the early 1900s (McCarthy, 1954), the studies in the area of computer-
mediated writing (cf. Baron, 2004) have also pointed out differences between males and
females: girls appear to use SMS messaging more often than boys, writing longer and more
complex messages with more "emotional" content (Thurlow & Poff, in press). Helsper (2010)
improves this idea with a model proposing that generation determines the level of Internet
use and life stage determines gender differences in Internet use. Herring (2000) has
emphasized that CMC reproduces the same dominant male/female rapport which has been
observed in traditional interactions: men control public discourse (such as forums, for
example) while women tend to interact within the private sphere in smaller groups.
practice) of CMC in general, or of another specific medium, has an influence on what the
writers produce. Beach and Lundell (1998) have shown how CMC experience among 12- and
13-year-old middle school students actively engaged them in reading and writing, seen as
“novice” and “advanced” users in a variety of online communities can be made using
structural characteristics or discourse pragmatics. The use of a certain medium enables the
creation (in a virtual space) of an identity as well as of linguistically marked social relations.
Currently, there is very little data concerning the French language, so the results of our
study will enlarge the available knowledge in this field by adding data from a new language.
Surveys on French-speaking populations have pointed out the massive use of SMSes by
adolescents. A study by Leo and Wolf4 (2004), and another by Samsung (2006) in Belgium,
showed that 75% of the Belgian population has a mobile phone. Furthermore, 90% of the 15-
4
Leo and Wolf is a company specializing in mobile marketing.
13
to 35-year-olds use the text-messaging function. According to the CREDOC survey conducted
in France (Bigot & Croutte, 2010), the proportion of mobile-phone users rose from 24% in
1999 to 83% in 2010 (84% of the 12- to 17-year-olds). In 2003, 57% of the mobile-phone
owners used the SMS function; in 2010 this figure reached 73% (97% of the 12- to 17-year-
olds). The average weekly number of SMS messages was 57 (182 for the 12- to 17-year-olds).
Crystal (2008) and Thurlow and Poff (in press) found comparable figures for various
languages (Roman alphabet-based or otherwise) and countries including the United States,
the United Kingdom, Finland, Sweden, Norway, Germany, Italy, South Africa, Nigeria, New
Zealand, Kuwait, Malaysia, Japan, Korea, China, Taiwan, and Hong Kong.
4. Method
Two research centers at the Catholic University of Leuven, Belgium, the CENTAL
(Center for Automatic Language Processing) and the CELEXROM (Center for Studies on
SMS Messages to Science"5 aimed at compiling a corpus of SMS messages for research in
linguistics and applied linguistics. To participate in the project, potential informants had to
corpus containing 30,000 SMS messages from 2,436 informants (between the ages of 12 and
73) was compiled for the project (Fairon et al., 2006). The data was collected between
4.1. Participants
5
http://www.sms4science.org/
14
From the entire "Donate Your SMS Messages to Science" corpus, we selected 91
participants on the basis of the three following criteria: language (participants had to be
French-speaking only), age (participants had to be between 15 and 18 years of age) and SMS
was studied using three variables. Gender was used, together with age, to establish 8
a) Age, two groups: 15-16 years (n = 41) and 17-18 years (n = 50)
c) SMS experience, in terms of the number of years and frequency of SMSing. We set
been filled out by the participants (cf. section 3.2. concerning procedure and table
Insert Table 1
4.2. Procedure
To participate in the project "Donate Your SMS Messages to Science", people had to
forward SMS messages of their choice to a free phone number. It must be stressed that,
particularly for an adolescent population, no matter the payment plan of the phone, the SMS
donation was free of charge. The individuals were also asked to fill in a sociolinguistic
6
We included in our sample all participants that satisfied our criteria except for some of the 17- to 18-year-olds
who were experienced SMS users: in order to balance the group sizes, only the first 35 of these 49 experienced
participants were included in the sample. The number of SMSes collected per participant varied between 1 and
59.
15
(language, age, gender, etc.) and experience in text-messaging. The questionnaire was
addressed to a very large audience (informants between the ages of 12 and 73), did not ask
questions about the type of payment plan (i.e., limited, unlimited), its price, or the person
paying it. However, with regard to the adolescents we retained within our sample, we
deducted, without great risk of error, that the payment plan was not unlimited (due to the
date of the collection—2004), that the parents were paying for the phone service, and that
the low cost, as compared to making phone calls, was one of the motivations behind the use
of SMSes. As for the “experience in text-messaging” variable in our sample, the experienced
participants had more than three years of text-messaging practice and claimed to send over
10 SMS messages per week (between 10 and 100). The inexperienced group had less than 3
years of text-messaging practice and claimed to send less than 10 SMS messages per week
(between 0 and 10). The experience variable was therefore defined by the two items
originating from the questionnaire: the length and the frequency of SMS use. Table 2 gives
the number of participants in each experience group, by the number of SMS messages the
Insert Table 2
For each participant, we retained all SMSes submitted. Table 3 indicates the number of
messages7 each, for age, gender, and SMS experience level (864 messages in total).
7
For the participants of both experience levels, the number of messages actually turned in was independent of
the number they claimed they sent per week.
16
Insert Table 3
4.3. Coding
We examined three types of message indexes: quantity indexes, structure indexes, and
function indexes.
For each original message, we used the automatic counting formulas in Excel® to tally
the number of characters (including spaces; the value of this index varied by definition
between 1 and 160) and the number of words (defined as a letter string with a space on
Original message: Euh d'accord c'est gentil! Je serai vers 7h55 devant chez toi! Bisous bonne
soirée! (Uh ok, cool. I’ll be outside ur place @ 7:55! Bye hava good 1).
Original message: Tit mess pour savoir comment tvas depuis ltemps! (a quick message t'find
Each message was broken down into three segments: the opening (O), the message
itself (M), and the closing (C). These three segments correspond to the structure defined by
Herring (1996b) regarding written language and used to study the Usenet and LISTSERV
messages. Our objective was to determine whether SMS users respect this three-phase
17
structure: they greet each other (O), they say what they have to say (M), and they take leave
(C). In interactions taking place via SMS, unlike traditional oral (Goffman, 1967) or written
exchanges (Adam, 1998; Herring, 1996b), there are messages with no opening, no closing, or
neither of the two. We defined four types of messages (see below with examples).
Message alone (M). M: Ya aussi p34 et 56. Tu dois aller chercher tt ce ki ya sur moodle
en anglais. (There are pages 34 and 56 too. You have to go look for everything there is
Opening and message (O+M). O: Salut tu vas bien? (Hi, how're you doing?) + M: G
pRdu livr d1 pot. Il è pa ché toi? 7 1 sur mart1 lutR king. (I lost a friend of mine's book.
Message and closing (M+C). M: Ben tanto oui mai po mtn jai ecol bizoux (Well, this
afternoon yes, but not now. I have school.)+ C: a tanto jtad (See you this afternoon.
Love you)
Opening, message, and closing (O+M+C). O: Salut (Hi) + M: keske tfé c soir?11srevoi
kan (What are you doing tonight? When are we going to see each other again?) + C:
CIAO (BYE)
A coder-agreement index for structures was defined as being equal to the number of
structures coded in the same way, divided by the total number of messages coded. In all, 101
messages (chosen randomly) were coded by two coders. The mean coder-agreement index
was 94%.
18
As Thurlow and Brown (2003) noted for SMSes written in English, text messages have
transactional and relational. This was the classification we chose, by adapting it to our data.
personal favors. During the coding process, for the sub-categories a) and b), the inter-coder
agreement rate was less than 50%: In fact, an exchange of practical details can also be
interpreted as a request for a personal favor. For this reason, we chose not to have a sub-
category for the informative-transactional function. From the relational function category,
friendly relations, and love relations. One and the same message could fall into several
functions was calculated in the same way as the coder-agreement index for structures: the
mean coder-agreement index here was 88.3%. As Thurlow and Brown (2003) emphasize:
information exchange for analytical convenience, it can be revealing to compare the relative
responses” (p. 10). The coding was carried out in terms of the main orientation of the
message.
19
Informative-transactional function
favor done for the interlocutor. jtapporte le livr dmain (I’ll bring you the book
tomorrow).
Requests for personal favors but also resquest for information. si t as 1 2ème
calculette tu pe la prendre?g oublié la mienne (If you have another calculator could you
Relational function
Ritualized interpersonal exchanges. coment ale vous koi d9 (how are you what's
new)
Social arrangements. 2m1, rdv place keym a 8h. (tomorrow, meet at Place Keym at 8
a.m.)
Friendly relations. jsuis 2 tt coeur heureuse pr toi! Jte souhaite tt le bonheur que
qq1 d'aussi exceptionnel que toi (I'm truly happy for you! I wish you all the happiness
Love relations. j'ai pas eu mon petit bonne nuit hier mais c'est pas grave!Tu peux
pas imaginer comme tu es importante pour moi mon tit bébé!! (I didn't get my little
good night last night but it's ok! You can't imagine how important you are for me my
little baby!)
In the findings presented here, we did not analyze in great detail the different sub-
categories of each relational function category. We focused on the analysis of our study
20
objective: to compare, as did Thurlow and Brown in 2003, the proportion of messages
a relational function.
Table 4 provides a synthesis of the factors and dependent variables which we used to
analyze our results. With the SMSes as the random variable, ANOVAs were performed upon
Insert Table 4
5. Results
We examined three types of message indexes: quantity indexes, structure indexes, and
function indexes.
The data was processed in an ANOVA with three factors: age (15-16 and 17-18 years),
Number of characters (including spaces): Table 5 gives the mean number of characters
as a function of age and gender. The effect of gender was significant (F(1,856) = 51.11,
p < .00001, ² = .06), as was the age-by-gender interaction (F(1,856) = 24.38, p < .00001, ² =
.03). As a whole, girls produced more characters than boys did. The difference between girls
21
and boys was significant among the 15- to 16-year-olds (F(1,856) = 56.71, p < .00001; 129
characters for girls and 82 for boys) and nonsignificant among the 17- to 18-year-olds (about
Insert Table 5
Number of words: Table 5 also gives the mean number of words as a function of age
and gender. The effect of gender was significant (F(1,856) = 28.91, p < .00001, ² = .03), as
was the age-by-gender interaction (F(1,856) = 29.86, p < .00001, ² = .03). As a whole, girls
wrote more words than boys did. The difference between girls and boys was significant
among the 15- to 16-year-olds (F(1,856) = 45.62, p < .00001; 24 words for girls and 16 for
boys) and nonsignificant among the 17- to 18-year-olds (about 21 words per message).
Four message structures were considered: message alone (M), opening and message
(O+M), message and closing (M+C), and opening, message, and closing (O+M+C). In all, 73%
(N = 630) of the total number of messages (N = 864) did not have the traditional O+M+C
structure, which means that a mere 27% (N = 234) of the messages did have this structure.
In cases where the structure was not traditional: structure M, where the message
stood alone without an opening or closing, represented 23.4% (N = 202) of the messages,
structure M+C represented almost 47% (N = 406) of the messages, and structure O+M had a
very low frequency (2.6%, N = 22). Overall, 30% of the messages had an opening, 74% had a
The structure indexes were processed in the same way as the quantity indexes. For the
three most frequent structures (M, M+C, and O+M+C), we used an ANOVA to analyze the
number of occurrences per message as a function of age, gender, and experience. For type
M+C, no factor was significant. For type M (see Table 6), the effects of gender
(F(1,856) = 7.08, p < .01, ² = .008) and experience (F(1,856) = 6.52, p < .01 , ² = .007) were
significant, as were the age-by-gender interaction (F(1,856) = 5.56, p < .01 , ² = .006) and
whole, type M was produced more often by boys than by girls, and more often by
favor of the boys, more so at the age of 15-16 than at 17-18. Moreover, this age-by-gender
Insert Table 6
For structure O+M+C (see Table 6), the experience effect was significant
(F(1,856) = 5.57, p < .01, ² = .008). As a whole, type O+M+C was produced more often by
inexperienced participants was present only for 15- to 16-year-old boys (F(1,856) = 4.99,
p < .02), and 17- to 18-year-old girls (F(1,856) = 6.07, p < .01).
23
function with those that had a relational function. The data were processed in a four-factor
ANOVA: three between-group factors (age, gender, and experience) and one within-group
factor (message function: informative-transactional vs. relational), with the SMSes as the
random variable. We concentrated on the function effect and its interactions with other
factors (see Table 6). The function effect was significant (F(1,856) = 6.15, p < .01, ² = .007),
as were the function-by-age interaction (F(1,856) = 4.34, p < .03, ² = .005), the function-by-
gender interaction (F(1,856) = 5.29, p < .02, ² = .006). On the whole, the proportion of
the sum of the two proportions was greater than one since the same message could have
both functions. The interactions showed that the difference between the two functions was
more pronounced for the 15- to 16-year-olds than for the 17- to 18-year-olds and for the
girls than for the boys. Like Thurlow and Brown (2003), we calculated the percentages of
each function out of the total number of function occurrences. We obtained 43.57% for the
informative-transactional function (430 out of 987) and 56.43% for the relational function
6. Discussion
situations. The SMSes were collected using a reliable methodology that avoided transcription
24
errors and over-representation of short messages. The resulting data provides answers to
What were the characteristics of text messages from the point of view of length? How
did this index vary according to the message writer's sociolinguistic characteristics?
In terms of SMS length, our results (on average, 100 characters and 20 words) are
closer to those which were obtained from a small group of participants known to the
researchers (Deumert & Masinyana, 2008; Grinter & Eldridge, 2001; Thurlow & Brown, 2003)
than to those obtained via a phone survey with a large number of participants (Ling, 2005a,
2005b). An ecological collection demonstrates that between the ages of 15 and 18,
adolescents use between half and slightly more than three-fourths of the 160 characters
permitted by this form of technology. Moreover, our findings provide the first evidence of
changes according to age and gender: the girls' messages were indeed longer than the boys',
but only among 15- and 16-year-olds. The girls and the boys seemed to have different
developmental paths: between the ages of 15-16 and 17-18, the girls' message length
(already high) stayed the same, whereas the boys’ message length increased to become
equivalent to that of the girls. These findings offer a different slant on Ling's study (2005a,
2005b) showing that messages produced by females were longer and more complex than
those of males (between 13 and 34 years of age). However, it is difficult to compare our data
with that of Ling (2005a, 2005b) due to the different method of ecological collection/survey
which was used and the scale of the variations: in our study, the difference between boys
and girls ranged from 80 to 130 characters, while those of Ling (2005a, 2005b) only ranged
from 20 to 30. As Berko-Gleason stressed, “Most of the gender differences that do exist in
boys’ and girls’ language are more likely to be the product of socialization and context than
25
the result of innate biological differences.” (p. 416). From a social point of view, the
precociousness of girls may be explained by the fact that with CMC, just as in traditional
interactions, the female gender tends to put a priority on the private sphere and smaller
groups (Herring, 2000). This corresponds to the reality of SMS messages which are
essentially meant for close friends or family members (Ischii, 2006; Reid & Reid, 2005;
Rivière & Licoppe, 2005). It should be mentioned that, while the differences between girls
and boys disappeared around 17-18 years of age in our corpus of French-language SMSes
collected in 2004, these differences still exist in other online environments (Kapidzic &
What were the characteristics of text messages from the structural point of view? How
did this index vary with the message writer's sociolinguistic characteristics?
Regarding the structure indexes, 73% of the messages did not follow the mandatory
the messages had an opening, 74% had a closing. Our results are consistent with those which
have already been presented by Ling (2005b), Spagnolli and Gamberini (2007) and Laursen
proportion of SMS messages exist without an opening and/or a closing. This proportion is
greater than that found for another type of technically-mediated communication: the email
out that the proportion of SMS messages considered “non-canonic” varies from one study to
the next. The very high proportion (90%) which resulted from Ling’s (2005b) phone survey
messages and thus, of messages without an opening and/or a closing. In addition, Spagnolli
26
turns, with an opening on the first message and a closing on the last. However, this
phenomenon cannot explain why, in our study, 74% of the messages had a closing while only
30% had an opening. If Spagnolli and Gamberini's (2007) hypothesis were correct, the
opening and closing percentages would be below 100%, but equal. These authors
emphasized that with SMSes, salutations are made using other non-verbal methods (the
arrival of the message is announced by various alerts, the identity of the sender is displayed
by the device) and are therefore not necessary. As the end of the exchange is not signaled by
technical means, it is more often shown by a form of closing. Spilioti (2011) offers an
Of course, an SMS exchange is a kind of interaction that allows for the absence of a closing
since it is not necessarily expected by the interlocutors. However, the presence of a closing
denotes a special social situation where politeness (in terms of respecting a ritual) is
necessary for the interlocutors to save face, such as in the case of social distance,
With respect to prior studies, our data provides new information about the role played
appeared in 15- to 16-year-old boys, the group which, in terms of development, took on SMS
messaging the slowest. Only long-term, frequent practice enabled these boys to enter into a
conversations involving several speaking turns (Spagnolli & Gamberini, 2007). In sum, our
data confirm that a difference does indeed exist between boys and girls. But our results
indicate that this difference is mitigated on the one hand by age, and, on the other hand, by
experience.
What were the characteristics of text messages from the point of view of the function?
How did this index vary with the message writer's sociolinguistic characteristics?
Regarding the functions of text messages, our results are consistent with Thurlow and
Brown's (2003) and with all earlier studies showing the importance of the relational function
of SMSes. Moreover, for both the informative and relational functions, we validated a coding
grid that gave us a high between-coder agreement rate, something that had not been done
in past research. In line with our quantity indexes and structure indexes, we showed here
relational functions only prevailed for 15- to 16-year-olds, and for girls.
Our study enabled us to show that the SMS writers' characteristics (age, gender, and
SMS experience) played a role in the length, structure, and function of the text messages
they wrote. Long messages were mainly produced by 15- to 16-year-old girls, whose
messages were also largely relational. The experienced 15- to 16-year-old boys generally
wrote messages without an opening or a closing and produced few traditional messages with
data collection should be extended to include both younger adolescents (ages 11-12 and 13-
A next step could be to compare the distribution of informative (44%) and relational
(56%) functions in SMS messages, with the distribution observed for other types of
technically-mediated communication such as oral communication over the phone (land lines
or cell phones), oral communication via video chatting, written communication via instant
messaging, online help forums, and written communication on social networks (e.g.,
Facebook, Twitter).
In the long run, relating the functions of SMS messages to their linguistic
characteristics (length and dialogical structure, but also orthographic forms) is likely to be
that which will provide the greatest amount of insight into the SMS register.
containing 30,000 SMS messages from 2,436 informants (between the ages of 12 and 73). A
adolescents with SMSes produced by adults. A comparison between the French language
and others languages will require the completion of studies in different countries using the
same methodologies.
particularly studies that take the speaking-turn sequence of the two writers into account,
and ones that look at the relationship between the functions and linguistic forms of text
messages. Our data bring out the complexity of the SMS register. They argue in favor of
somewhere between the oral and written registers, but as a register of its own that differs
29
from traditional oral communication, traditional written communication, and sign language.
Unlike these other three major registers, computer-mediated writing does not remain stable
over time8 and currently, its extremely rapid rate of change poses a true challenge to
Acknowledgments
This research was supported by the French National Research Agency project ANR-08-
and Learning". Special thanks are extended to Vivian Waltz and Martha Randy for translating
this paper.
8
The first analyses of the francophone corpus collected as part of the sms4science project between 2004 and
2010 showed diachronic differences: e.g., the messages seemed to be shorter and shorter (Fairon, 2011).
30
Questionnaire SMS4Science
AUTHORS’ PERSONAL PROFILE
(Closed-choice questions)
What is your age?
What is your gender?
What is your field of work (travel agent, librarian, salesperson, teacher, engineer, soldier,
secretary, etc.)?
What is your current level of study or highest diploma if you have finished your studies
(primary, general secondary, technical secondary, superior professional secondary, etc.)?
What is your native language (German, English, Arabic, Italian, Luxembourgish, etc.)?
What other languages do you speak fluently (German, English, Arabic, Italian,
Luxembourgish, etc.)?
What is the postal code of your residence?
What is your country of residence (Belgium, France, Luxembourg, etc.)?
How many keys does your mobile phone have (min./max.)?
Does it have a dictionary (yes/no)?
Rank the addressees of your SMSes by order of preference (never, rarely, regularly, almost
daily):
Never Rarely Regularly Almost daily
Family
Friends
Colleagues
Contests/games
What other modes or systems of communication on the Internet do you use? (never, rarely,
regularly, almost daily)?
Never Rarely Regularly Almost daily
SMS from a website
Internet chat
Internet forums
Email
32
References
pragmatique des pratiques discursives. In J. Siess (Ed.), La lettre entre réel et fiction
Anis, Jacques, 2007. Neography: Unconventional spelling in French SMS text messages In :
Danet, B. and Herring, S. C. (Eds), The multilingual internet: Language, culture and
Austin, John L., 1962. How to do things with words. Cambridge, Massachusetts, Harvard
University Press.
Baron, Naomi S., 2004. See you online: Gender issues in college students' use of instant
Beach, Richard & Lundell Dana Britt, 1998. Early adolescents' use of computer-mediated
communication in writing and reading. In D. Reinking, M.C. McKenna, L. Labbo & R.D.
Berko Gleason, Jean. 2001. The development of language. Needham Heights: Allyn and
Berman, Ruth A., 2005. Introduction: Developing discourse stance in different text types and
Bernicot, Josie, 1994. Speech acts in young children: Vygotski’s contribution. European
Bernicot, Josie & Mahrokhian, Alénouche, 1989. To ask and to insist after a refusal: how do
Bernicot, Josie, Comeau, Judith & Feider, Helga, 1994. Dialogues between French-speaking
mothers and daughters in two cultures: France and Quebec. Discourse Processes, 18,
19-34.
Bernicot, Josie, Laval, Virginie & Chaminaud, Stéphanie, 2007. Nonliteral Language Forms in
Children: In What Order Are They Acquired in Pragmatics and Metapragmatics? Journal
Bigot, Régis, & Croutte, Patricia, 2010. La diffusion des technologies de l'information et de la
Bou-Franch, Patricia, 2011. Openings and closings in Spanish email conversations. Journal of
Brown, Alexandra, 2002. The language and communication of SMS: An exploratory study of
Crystal, David, 2001. Language and Internet. Cambridge, Cambridge University Press.
Crystal, David, 2008. Txtng: the gr8 db8. Oxford, Oxford University Press.
Crystal, David, 2010. The Cambridge Encyclopedia of Language (third edition). Cambridge:
Danet, Brenda, & Herring, Susan C., 2007. The multilingual internet: Language, culture and
Deumert, Ana, & Masinyana, S Sibabalwe Oscar, 2008. Mobile language choices. The use of
English and isiXhosa in text messages (SMS): Evidence from a bilingual South African
Döring, Nicola, Hellwig, K. & Klimsa, Paul, 2005. Mobile Communication among Youth in
Germany. In: Nyiri, K (Ed.), A Sense of Place: The Global and the Local in Mobile
Drouin, Michelle A, 2011. College students’ text messaging, use of textese and literacy skills
Ellison, Nicole B., Steinfield, Charles, & Lampe, Cliff, 2007. The benefits of Facebook
"Friends:" Social capital and college students' use of online social network sites. Journal
http://jcmc.indiana.edu/vol12/issue4/ellison.html
m.fr/IMG/pdf/Fairon_Cedrick.pdf
Fairon, Cedrick, Klein, Jean & Paumier S, 2006. SMS pour la science. Corpus de 30.000 SMS
Faulkner, Xristine & Culwin, Fintan, 2005. When fingers do talking: a study of text messaging.
Fernandez, Julieta & Yuldashev, Aziz, 2011. Variation in the use of general extenders and
Frehner, Carmen, 2008. Email – SMS – MMS: the linguistic creativity of asynchronous
discourse in the media age. Studies in Language and Communication, Vol. 58. Bern:
Peter Lang.
Goffman, Erving, 1967. Interaction ritual: Essays on face-to-face behavior. Garden City, NY,
Anchor.
35
Grice, Paul, 1975. Logic and conversation. In: Cole, P. & Morgan, J. (Eds), Syntax and
Grinter, Rebecca E, & Eldridge, Margery A, 2001. Y do tngrs luv 2 txt msg? In: Prinz, W., Jarke,
M., Rogers, Y., Schmidt, K. & Wulf, V. (Eds), Proceedings of the seventh European
Publishers.
Hård af Segerstad, Ylva, 2005. Language Use in Swedish Mobile Text Messaging. In: Ling, R.
London, Springer.
Harper, Richard, Palen, Leysia, & Taylor, Alex, (Eds.) 2005. The Inside Text. Social, Cultural
Helsper, Ellen. J, 2010. Gendered Internet Use Across Generations and Life Stages.
Herring, Susan. C., 1996a. Computer-mediated Communication: Linguistic, Social and Cross-
Herring, Susan. C., 1996b. Two variants of an electronic message schema. In Herring, S.C.
Herring, Susan. C., 2000. Gender differences in CMC: Findings and implications. Computer
http://cpsr.org/issues/womenintech/herring/
36
online behavior. In Barab, S.A., Kling, R. & Gray, J.H. (Eds.), Designing for Virtual
Communities in the Service of Learning (pp. 338-376). New York: Cambridge University
Press.
Herring, Susan. C., 2007. A Faceted Classification Scheme for Computer-Mediated Discourse.
(http://www.languageatinternet.org/articles/2007)
Höflich, Joachim, & Gebhardt, Julian, 2005. Changing Cultures of Written. Communication:
Letter - E-mail –SMS. In: Harper, R., Palen, L. A. & Taylor, A. (Eds), The inside text:
Social, cultural and design perspectives on SMS, Dordrecht, The Netherlands, Springer.
Ishii, Kenichi, 2006. Implications of Mobility: The Uses of Personal Communication Media in
Jisa, Harriet, Reilly, Judy. S., Verhoeven, Ludo, Baruch, Elisheva, & Rosado, Elisa, 2002.
Kapidzic Sanja & Herring, Susan C., in press. Gender, Communication, and Self-Presentation
Communication.
Kasesniemi, Eija L., 2003. Mobile messages: Young people and a new communication culture.
Kemp, Nenagh, 2011. Mobile technology and literacy: effects across cultures, abilities and
Laursen, Ditte, 2005. Please reply! The replying norm in adolescent SMS communication. In:
Harper, R., Palen, L. A. & Taylor, A. (Eds), The inside text: Social, cultural and design
Leo & Wolf, 2004. 70% des 15-35 ans acceptent de recevoir de la publicité sur leur téléphone
ans-acceptent-de,343+.html
Ling, Richard, 2005a. The length of text messages and the use of predictive texting: Who uses
it and how much do they have to say? Paper presented at the Association of Internet
Ling, Richard, 2005b. The socio-linguistics of SMS: An analysis of SMS use by a random
Carmichael’s manual of child psychology (2nd ed.). New-York, Wiley and Sons.
Noveck, Ira & Sperber, Dan, 2004. Experimental Pragmatics. Basingstoke, England, Palgrave.
Panckhurst, Rachel, 2009. Short Message Service (SMS): typologie et problématiques futures.
Montpellier 3.
Puustinen, Minna, Bernicot, Josie & Bert-Erboul, Alain, 2011. Written Computer-Mediated
Requests for Help by French-Speaking Students: An Analysis of Their Forms and Functions.
Puustinen, Minna, Volckaert-Legrier, Olga, Coquin, Danièle, & Bernicot, Josie, 2009. An
design of ecologically valid supporting tools. Computers & Education, 53, 1040-1047
Ravid, Dorit. & Tolchinsky, Liliana, 2002. Developing linguistic literacy: a comprehensive
Reid, Donna, J. & Reid, Frase, J.M., 2005. Textmates and text Circles: Insights into the Social
Ecology of SMS Text messaging. Mobile World Computer Cooperative Work, 2, 105-
118.
Rivière, Carole. A., & Licoppe, Christian, 2005. From voice to text: Continuity and change in
the use of mobile phones in France and Japan. In: Harper, R., Palen, L. A. & Taylor, A.
(Eds), The inside text: Social, cultural and design perspectives on SMS, Dordrecht, The
Netherlands, Springer.
Rosen, Larry D., Chang, Jennifer, Erwin, Lynne, Carrier, L. Mark & Cheever, Nancy A., 2010.
The Relationship Between “Textisms” and Formal and Informal Writing Among Young
Samsung, 2006. Le comportement mobile belge décortiqué par Samsung. Retrieved 22 April
decortique-par-samsung/
Savas, Perihan, 2011. A case study of contextual and individual factors that shape linguistic
Searle, John. R., 1969. Speech acts. Cambridge, MA, Cambridge University Press.
Searle, John. R., & Vanderveken, Daniel, 1985. Foundations of illocutionnary logic.
Spagnolli, Anna, & Gamberini, Luciano, 2007. Interacting via SMS: Practices of social
Spilioti, Tereza, 2011. Beyond genre: closings and relational work in text-messaging. In:
Thurlow , C. and Mroczek, K., (Eds.) Digital discourse: language in the new media (pp.
Strömqvist, Sven, E., Johansson, Victoria, Kriz, Sarah, Ragnarsdóttir, Hrafnhildur, Aisenman,
Ravid, & Ravid, Dorit, 2002. Towards a crosslinguistic comparison of lexical quanta in
Thurlow, Crispin, 2006. From statistical panic to moral panic: The metadiscursive
construction and popular exaggeration of new media language in the print media.
Thurlow, Crispin, & Brown, Alex, 2003. Generation Txt? The sociolinguistics of young people's
http://faculty.washington.edu/thurlow/papers/Thurlow (2003)-DAOL.pdf
Thurlow, Crispin & Poff, Michele, in press. The language of text messaging. In: Herring, S.C.,
Stein, D. & Virtanen, T. (Eds.), Handbook of the pragmatics of CMC. Berlin and New
Vignaux, Georges, 2003. Du signe au virtuel: les nouveaux chemins de l’intelligence. Paris :
Editions le Seuil.
40
Volckaert-Legrier, Olga, Bernicot, Josie, & Bert-Erboul, Alain, 2009. Electronic mail, a new
Waldvogel, Joan, 2007. Greetings and closings in workplace email. Journal of Computer-
Wood, Clare, Jackson, Emma, Hart, L., Plester, Beverly & Wilde, Lucy, 2011. The effect of text
Table 1.
Female 12 20 32
Total Experienced 25 35 60
Female 9 8 17
Total Inexperienced 16 15 31
Total 41 50 91
42
Table 2.
Experienced participants
11 to 20 21 to 50 51 to over 100
messages per messages per messages per
week week week
Number of experienced
33 18 9
participants
Inexperienced participants
Less than 5
messages per 5 to 10 messages per week
week
Number of inexperienced
10 21
participants
Note: The experienced participants had more than 3 years of text-messaging practice and claimed to send over 10 SMS
messages per week (between 11 and 100). The inexperienced group had less than 3 years of text-messaging practice and
claimed to send 10 or fewer SMS messages per week (between 0 to 10).
43
Table 3.
Inexperienced Male 22 41 63
group
Female 50 94 144
Table 4.
Factors
Levels
Experience Experienced/Inexperienced
Dependent variables
Values
QUANTITY INDEXES
Table 5.
Mean Number of Characters per SMS Message and Mean Number of Words per SMS
Table 6.
STRUCTURE INDEXES
Proportion of messages with a message-alone structure (M), by participant age, gender, and SMS experience
15-16 17-18
Male Experienced group 0.44 (0.04) 0.21 (0.03)
Inexperienced group 0.14 (0.09) 0.29 (0.06)
Total Male 0.29 (0.05) 0.25 (0.03)
Female Experienced group 0.11 (0.03) 0.31 (0.03)
Inexperienced group 0.10 (0.06) 0.17 (0.04)
Total Female 0.10 (0.03) 0.24 (0.03)
Proportion of messages with an "opening-message-closing" structure (O+M+C), by participant age, gender and
SMS experience
15-16 17-18
Male Experienced group 0.13 (0.04) 0.28 (0.03)
Inexperienced group 0.36 (0.09) 0.24 (0.07)
Female Experienced group 0.36 (0.03) 0.20 (0.03)
Inexperienced group 0.36 (0.06) 0.34 (0.04)
FUNCTION INDEXES
Proportion of messages with each function (informative-transactional or relational) in the body of the message,
by participant age, gender, and SMS experience.
15-16 17-18
Informative-transactional 0.52 (0.03) 0.56 (0.02)
Relational 0.71 (0.03) 0.57 (0.02)
Male Female
Informative-transactional 0.58 (0.03) 0.50 (0.02)
Relational 0.59 (0.03) 0.69 (0.02)
Note: Standard deviations are shown in parentheses, and the highest significant values are indicated in bold print.
47
Josie Bernicot is a professor at the University of Poitiers (France). Her research deals with
children’s language acquisition. In addition to typical and atypical oral language in children,
her studies also focus on the use of new communication technologies by adolescents. In
France, she has contributed to the development of the pragmatic perspective and the study
of the social uses of language. She is the leader of a focus team on this subject at the
Cognition and Learning Research Center (CeRCA-CNRS), as well as a member of both the
Olga Vockaert-Legrier received her Ph.D. from the University of Poitiers (France). Currently,
its linguistic features, especially email and SMS messages as a new register of written
language.
Antonine Goumi received her PhD in cognitive psychology from the University of Poitiers
(France). She began her work as a researcher at the Cognition and Learning Research Center
(CeRCA-CNRS) and now works at the National Center for Pedagogical Documentation, under
the supervision of the French Ministry of Education. Her research interests are related to
technologies for education, the role they play in teaching at school and in the evaluation and
improvement of learning.
Alain Bert-Erboul is a researcher at the CNRS (French National Research Agency) and works
at the Cognition and Learning Research Center (CeRCA) at the University of Poitiers (France).