motivations for the use of “smilies” as discourse markers in twitter messages and their effect...
DESCRIPTION
Short essay about the use of Smilies / or emoticons in Twitter messages.TRANSCRIPT
Motivations for the use of “Smilies” as discourse markers in Twitter messages and their effect on the
frequency of co-occurring pronominal phrases
Introduction
Emoticon use has grown rapidly over the past 10-15 years in correlation with the explosion of online
communications and sms text messaging. What started as a niche habit, associated with other forms of ‘text-
speak’ and ‘net-lingo’ is now fully entrenched into the mainstream and can be found readily across all
aspects of written computer-mediated-communication. From offices to chat-rooms, the emoticon has become
sedimented and normalised as a discourse marker used to indicate a variety of emotional affect including
irony, humour, anger, solidarity or pleasantry. Twitter is the relative ‘new-kid-on-the-block’ and is
multimodal, acting as public broadcasting, blogging and IM service.
Aims
The aim of this study is to shed light on the kind of linguistic environment in which emoticons occur. At the
most basic level, for example, the presence of a smiley “:)” is likely to coincide with an increased likelihood
of language associated with positive emotion. The data presented here shows also; gratitude, the sweetening
of requests (“follow” / “please”) and interestingly, use of the 2nd person singular “you” and to a lesser
degree its shortened counterpart “u”.
Literature Review - Motivations
Interviews conducted by Shortis (2013), have shown repeatedly that people we might bracket as normative
in their approach to language use, who consider reworking of spelling in new media as an irritating
obfuscation, still widely consider the emoticon as an important discourse marker, and in many ways an
addition to the punctuation set. Those who value highly clarity of expression, endorse emoticon use across a
variety of media and social context.
Provine et al (2007) describe emoticons as ‘punctuating website text messages’ in a manner comparable to
the way ‘laughter “punctuates” speech’ and specifically casual speech;
“Emoticons are an unnecessary and unwelcome intrusion into well-crafted text. But emoticon use is
better contrasted with colloquial speech than formal writing or literature” (Provine et al., 2007).
Method
Two corpora were gathered from Twitter.com by continually scrolling down the page to force it to keep
displaying new tweets spanning back in time and then copy/pasting the resulting list of tweets into a plain
text document. This process was performed four times for each corpus, each time for about 10 minutes. Each
resulting corpus contains approximately 150,000 words. All_Eng corpus contains all tweets composed in
Exam number: Y0097746
English whilst the SMILE corpus contains all tweets composed in English that include the emoticons :), :-), :D
and :P. The corpuses were then put through Wordsmith Tools, to generate Wordlists and concordances. This
data was then put alongside a wordlist of the most frequently occurring words in Twitter messages compiled
by the OUP (Oxford University Press) based on 1.5 million tweets, to validate the rates that were being
shown from All_Eng corpus. On the whole, the results were comparable.
Results
The table below shows the proportion of each smily variant in a corpus of 150,000 words (Approx. 9.5k
Tweets) filtered to contain only tweets that used some emotionally positive smiley variant ( :( and :-( not
included). These rates are comparable to other findings in the new-media literature, with :) acting as the
default, go-to, smiley, :D for emphasis, :-) more elaborate and :P to denote sarcasm.
Smiley Frequency Percentage
:) 6527 68.4%
:D 1461 15.3%
:-) 741 7.8%
:)) 413 4.3%
:P 396 4.2%
Total 9538 100.0%
Table 1: Percentages of Smiley variants
Dresner and Herring (2010) state that the emoticon ‘can be viewed as an expansion of the text in the same
way that, for example question marks and exclamation marks are’. In All_Eng Corpus the smily face occurs
with about twice the frequency of the exclamation mark (291/124).
Similarly, the majority of smilies occur at the end of a tweet with 4087 of the 6527 tweets with :) having
nothing at all following it with approximately another 1000 followed by further finalisation of utterence (‘x‘,
‘xx‘, #name etc). So roughly 75% occurring at the end of a phrase, interestingly not as many as Provine et al
(2007) who found 99% emoticons occur at end of phrase across more general website texts. The smiley
represents roughly 0.2% of all tokens used on Twitter. What is interesting is that this ratio it fairly constant
whether a corpus is filtered to contain only tweets with non-standard spellings or is open to all English
tweets.
Exam number: Y0097746
(1) Seems like you just don't like me anymore :')
(2) actually a lot of people hate me lol :-) :-) :-)
(3) and idk i want to become better friends with you :)
(4) @name follow back. Thanks :)
(5) At least work was fun today :)
(6) Going to the park with the children, seems like a nice day :)
(7) very good thank you ;-) how's yourself name?X
Table 2: examples of Smiley usage
The above examples demonstrate the kind of functions served by the smiley. They are representative of the
proportion of phrase-final versus phrase medial. In line with Dresner & Herring’s (2010) drawing of a
parallel between emoticons speech act theory, smilies mirror real-world smiles to some degree. They serve to
show that an utterance is intended jokingly as in (1) and (2), to indicate that happiness (3) and (6), to sweeten
a request (4), to disambiguate (5) - in this instance to show that the statement is not sarcastic and the ‘winky’
‘;)’ used here as a greeting, and to delineate between discourses.
Word OUP SMILE ALL_ENG RATIO: SMILE/
ALL_ENG
thanks 0.09% 0.31 0.07 0.23
please 0.03% 0.28 0.08 0.28
follow 0.03% 0.43 0.14 0.32
birthday 0.02% 0.15 0.06 0.38
good 0.20% 0.43 0.17 0.40
awesome 0.04% 0.05 0.02 0.46
happy 0.05% 0.22 0.10 0.47
love 0.12% 0.47 0.25 0.53
hope 0.05% 0.11 0.06 0.55
cool 0.04% 0.05 0.03 0.59
see 0.10% 0.17 0.10 0.61
would 0.09% 0.12 0.08 0.66
you 0.75% 1.66 1.10 0.66
u 0.07% 0.27 0.19 0.71
going 0.13% 0.10 0.09 0.82
your 0.20% 0.37 0.31 0.83
are 0.25% 0.28 0.24 0.84
me 0.35% 0.76 0.65 0.86
so 0.26% 0.43 0.37 0.86
have 0.33% 0.34 0.30 0.89
lol 0.11% 0.19 0.18 0.92
Table 3: Lexical items showing increased frequency in SMILE corpus from most increased to least with pronominal phrases highlighted
Exam number: Y0097746
This table shows a marked increase in the frequency of 2nd person singular/plural as well as possessive.
Interestingly the 1st person singular ‘me’ shows and increase, whilst as shown in the table below, the 1st
person singular “I” shows little or no change. Perhaps this might point to smilies being used to make the
dialogue more ‘chat-like’. Tagliamonte & Dennis (2008, p.13) report that “first- and second-person pronouns
are more frequent in speech than in writing” and that earlier research from Yates (1996) on CMC at computer
conferences reflected this. With ‘you’ referencing the person with whom a speaker is interacting, it is perhaps
not a surprise that an increase in simulated outward gestures might be seen.
Word OUP SMILE ALL_ENG RATIO: SMILE/ALL_ENG
do 0.33% 0.22 0.22 0.99
it 0.76% 0.57 0.56 0.99
but 0.24% 0.23 0.23 1.00
and 0.85% 0.74 0.74 1.00
she 0.06% 0.08 0.08 1.03
for 0.71% 0.57 0.58 1.03
i 1.79% 1.71 1.77 1.04
be 0.29% 0.37 0.38 1.04
my 0.59% 0.67 0.70 1.04
just 0.32% 0.28 0.30 1.06
Table 4: Lexical items showing no change in SMILE corpus with pronominal phrases highlighted.
Word OUP SMILE ALL_ENG RATIO: SMILE/ALL_ENG
a 1.33% 1.08 1.19 1.11
with 0.36% 0.26 0.30 1.14
to 1.56% 1.12 1.33 1.19
on 0.60% 0.36 0.43 1.22
this 0.30% 0.30 0.37 1.23
like 0.20% 0.23 0.29 1.25
that 0.49% 0.31 0.39 1.25
omg 0.02% 0.04 0.05 1.27
funny 0.02% 0.03 0.04 1.32
out 0.23% 0.13 0.18 1.37
the 1.82% 1.10 1.51 1.37
not 0.25% 0.19 0.27 1.38
he 0.11% 0.07 0.10 1.38
is 0.83% 0.46 0.66 1.43
in 0.76% 0.45 0.67 1.51
of 0.75% 0.39 0.60 1.55
up 0.22% 0.16 0.26 1.61
at 0.37% 0.18 0.28 1.62
bad 0.04% 0.03 0.05 1.89
why 0.08% 0.06 0.11 1.89
they 0.13% 0.06 0.12 1.92
Table 5: Lexical items showing decreased frequency in SMILE corpus from most increased to least with pronominal phrases highlighted
Exam number: Y0097746
The above two tables show a small but clear preference for not using smilies in combination with masculine
or neutral pronouns ‘he’ and ‘they’, while ‘she’ is unchanged. Perhaps people actively avoiding combining
emoticons with male pronouns or maybe 3rd person pronouns are generally reduced by the use of
emoticons, with the female form being boosted back up to normal levels by association of their feminine
characteristic (Tossell et al., 2012).
Conclusion
Much is beyond the scope of a 1000 word paper and this is really a snapshot of a much larger view.
Questions that beg answering are; to what extent are online texts marked by their absence of discourse
markers? In which linguistic situations does a pressure to include some form of emoticon arise, how
widespread is this pressure and in such situations, how far can a decision to avoid their use be interpreted as
a deliberate veering away from a perceived childish or immature behavior as Tagliamonte (2008) describes.
Word Count = 1070
Exam number: Y0097746
References / Bibliography
Bays, H. (1998). Framing and face in Internet exchanges: A socio-cognitive approach. Linguistik online, 1(1).
Derks, D., Bos, a. E. R., & Von Grumbkow, J. (2007). Emoticons and Online Message Interpretation. Social
Science Computer Review, 26(3), 379–388.
Dresner, E., & Herring, S. C. (2010). Functions of the Nonverbal in CMC: Emoticons and Illocutionary Force.
Communication Theory, 20(3), 249–268.
Tossell, C. C., Kortum, P., Shepard, C., Barg-Walkow, L. H., Rahmati, A., & Zhong, L. (2012). A longitudinal
study of emoticon use in text messaging from smartphones. Computers in Human Behavior, 28(2), 659–663.
Provine, R. R., Spencer, R. J., & Mandell, D. L. (2007). Emotional Expression Online: Emoticons Punctuate
Website Text Messages. Journal of Language and Social Psychology, 26(3), 299–307.
Shortis, Tim (2013). Personal email correspondence.
Tagliamonte, S., & Denis, D. (2008). Linguistic Ruin? LOL! Instant Messaging and Teen Language. American
speech, 83(1), 3–34.
Exam number: Y0097746
Appendix
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
they
why
bad at
up
of
in
is
he
not
the
out
funn
y
omg
that
like
this
on
to
wit
h a
just
my be i
for
she
and
but it
do
lol
have
so
me
are
your
goin
g u
you
wou
ld
see
cool
hope
love
happ
y
awes
ome
good
birt
hday
follo
w
plea
se
than
ks
ALL_ENG
SMILE
!"#$%&"''(%$)'$&*)&+%$,$)'$&"-&$."/'")
0$12/3$14&)"&'52)6$&*)&+%$,$)'$&"-&$."/'")
7)'%$2,$8&"''(%$)'$&*)&+%$,$)'$&"-&$."/'")
Figure 1: 100% stacked graph showing the ratio formed by dividing the percentage of token occurrence in SMILE by ALL_Eng.
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
1.80
2.00
awes
ome
cool
funn
y
omg
bad
birt
hday
hope
than
ks
plea
se
she
wou
ld
goin
g he
see
happ
y
why
they
follo
w
good
lol
out u do
but
are
love
up
not at
like
just
wit
h
have
your
this
so
be
that
on it
for of
me is
in
my
and
you a to
the i
SMILE
ALL_ENG
Figure 2: Line chart displaying token frequency as percentage of total word count in SMILE and ALL_Eng
Exam number: Y0097746