Monday, July 18, 2011

Communicating Meaning Online: A Digital Expression of Theory of Mind

© Creative Commons Attribution 2.0
Note: This post originally appeared on Scientific American.

The growth of email, instant messaging, texting, and various other digitally-mediated communicative tools (DMC) has been rapid and pervasive. The majority of people today are comfortable enough to use these communicative tools on a daily basis, particularly among younger generations. DMC appears to be a preferred means of communication. But the popularity of DMC forces us to deal with its main problem: How do these users clarify their meaning, intent, and desires?

The Demands of Social Life 

The ability to interpret social data is rooted in our theory of mind—our capacity to attribute mental states (beliefs, intents, desires, knowledge, etc.) to the self and to others. This cognitive development reflects some understanding of how other individuals relate to the world, allowing for the prediction of behaviors.1 As social beings we require consistent and frequent confirmation of our social placement. This confirmation is vital to the preservation of our networks—we need to be able to gauge the state of our relationships with others.

Research has shown that children whose capacity to mentalize is diminished find other ways to successfully interpret nonverbal social and visual cues 2-6, suggesting that the capacity to mentalize is necessary to social life. Digitally-mediated communication, such as text messaging and instant messaging, does not readily permit social biofeedback. However cyber communicators still find ways of conveying beliefs, desires, intent, deceit, and knowledge online, which may reflect an effort to preserve the capacity to mentalize in digital media.

The Challenges of Digitally-Mediated Communication

In its most basic form DMC is text-based, although the growth of video conferencing technology indicates DMC is still evolving. One of the biggest criticisms of DMC has been the lack of nonverbal cues which are an important indicator to the speaker’s meaning, particularly when the message is ambiguous.

Email communicators are all too familiar with this issue. After all, in speech the same statement can have multiple meanings depending on tone, expression, emphasis, inflection, and gesture. Speech conveys not only what is said, but how it is said—and consequently, reveals a bit of the speaker’s mind to interested parties. In a plain-text environment like email only the typist knows whether a statement should be read with sarcasm.

In the late 1980s and the early 1990s, the popularity of chat rooms grew dramatically. Individuals posting messages in these forums would attempt to be sarcastic or ironic and find that the absences of social cues resulted in their messages being taken quite seriously. This caused arguments and “flame” wars where the character of the offensive poster was dissected with derogatory comments sometimes resulting in the offender being banned from that chat room.7,8

To help resolve ambiguity, tech-savvy users turned to emoticons: “glyphs, usually representing stylized facial expressions, created mainly from short sequences of punctuation marks and are designed to convey an emotional tone in email.”7 The emergence of emoticons reminds us of the importance of interpreting social data in social life. We spend a great deal of our lives online—from shopping, to banking, to emails, to online gaming networks such as World of Warcraft, and virtual communities like SecondLife—and our need for a theory of mind remains unchanged.

Emoticons have been one solution to this problem. The website Emoticon Universe lists over 300 emoticons. These numerous symbols help us express ourselves online and suggest a need to make ourselves understood in the virtual arena.

© Creative Commons Attribution 2.0
Assessing Emoticon Use

If emoticons do convey differentiated and subtle meanings in the absence of readily available nonverbal and visual cues in text-based DMC, can they be used intentionally to manipulate group equilibrium? In informal discussions with high school and college students, and young professionals, the following themes emerged:
  • DMC provides an effective and easy method to keep in touch with people throughout the day, and lends itself easily to multi-tasking.
  • Emoticons are an accepted element of digitally-mediated communication.
  • Emoticons help users implicitly express tone and meaning, allowing for more nuanced text-based conversations.
According to one college-aged texter, “text messaging is simple—you can stay connected no matter where you go, but without smileys [emoticons] it would be easy to misinterpret things.” This sentiment is in keeping with research regarding why and how instant messaging is used by college students to maintain personal relationships: Participants mainly used IM to communicate with friends, and emoticons as well as emotext (Internet slang) are heavily integrated with IM.9 Emoticons provide users with a feedback loop, as illustrated by the following conversations copied from Facebook pages (grammar and spelling are preserved as they appear in the original postings).
Facebook Example 1:
Status: Duke won! Life is good!
Comment: You know, I’m beginning to find Facebook’s “like” feature very biased. If you can opt to give something a thumbs up, you should be able to give it a thumbs down. But until then …
Duke: Thumbs Down. =]
Emoticons help the writer express and check statements that could be unclear during the initial message production, but they also help the audience decipher cold text and understand the meaning the writer wishes to convey. In Example 1, someone posts a message expressing happiness that the Duke University basketball team has won a game. Although there are no emoticons present, the tone is clearly celebratory. The respondent’s comment is meant to disagree with the original poster, but not in a mean-spirited way. The addition of the smiley face emoticon at the end of the sentence indicates that the audience—and the owner of the original status—should not take the comment too seriously. If the comment had ended with a different emoticon the interpretation would have been different. For example, “Duke: Thumbs Down. >:/” gives the statement an angry intonation.
Facebook Example 2:
Status: Pisco and Meade have crossed over into a better place. RIP my beloved newts.
Comment 1: Noo ….. :*( I’m so sorry. Poor Pisco, poor Meade, I loved you guys!
Comment 2:  I had two little lizards a few years back and the same thing happened. :-[ I’m sorry.
Comment 3:  Oh, dear – I’m so sorry. I get weepy when my fish go, too – at least, the ones with personality.
Emoticons provide DMC users with a script for their responses. For example, seeing a sad face :-[ from a good friend may prompt you to ask how their day was, what happened, or perhaps let you know that you should give them a little space. In Example 2, the user posts that her pet newts have died. The responses include a crying emoticon (Comment 1) and a sad emoticon (Comment 2). Both comments are meant to sympathize with the original poster. Comments without sympathetic emoticons are still sympathetic (Comment 3), but the presence of an appropriate emoticon publically confirms the writer’s sympathy.

Resolving Ambiguity in Digital Media

With the rising popularity of microblogging sites such as Twitter, digital communicators must find ways to express thoughts and ideas both clearly and concisely.

Following the death of Michael Jackson, people turned to Twitter to share the news as well as their grief. Researchers from The Web Ecology Project conducted a sentiment analysis of Tweets to determine whether the people who tweeted about Michael Jackson’s death were less happy than normal. To assess this information, they collected statistical data via an ANEW analysis (Affective Norms for English Words) which revealed a lower valence score for Tweets concerning Michael Jackson as compared to everyday Tweets. Valence is the measure of pleasure versus displeasure; a higher score indicates greater pleasure.

Overall, the raters indicated approximately 75% of the 346 Tweets expressed sadness, but found a fair number of tweets to be difficult to interpret.10 Digital communication occurs between humans who are not equipped to statistically determine motive, intent, and desire. Humans require social cues. The difficulties the raters had in interpreting the tone of these statements strongly suggests the need for theory of mind tools—and is echoed in this study as interviewees expressed difficulty interpreting tone when social cues were absent as in the case of the third commenter remarking on the death of Pisco and Meade.

Emoticons provide social biofeedback and make text messages and IMs a little more personal. For example, without the benefit of a social prompt, “You’re late”  can be interpreted as serious/angry during virtual chat. The placement of an emoticon changes the tone of the statement: “You’re late ;p” implies that the speaker is joking, and the statement should not be taken seriously—similar to Example 1. Emoticons impart a bit of the user’s personality to the digital exchange--actually making it a conversation by employing differing intonations. These types of nuances are of importance because it allows relationships to be maintained over digital media.

Cultural Relevance in Digital Media

As DMC becomes the prevalent means of communication, it is being adapted to support and maintain our networks. This adaption is evident in the cultural modification of emoticons. For example, people from East Asia focus on eyes to distinguish between facial expressions, which according to researchers, demonstrates cultural specificity in expressing emotions.11 In a recent study, a group of Western Caucasian and a group of East Asian people observed faces that showed the following emotions: happy, sad, surprised, fearful, disgusted, angry, and neutral. The expressions were labeled in this fashion because they employed specific muscles that are associated with these emotions according to the Facial Action Coding System.

A review of emoticons popular in East Asian countries does reveal a tendency to vary “eyes” for expressions. For example, Western emoticons for happy include the following: : ) or : D. East Asian emoticons for happy are often depicted as follows: (^_^) or (^.^). In the second East Asian example, the mouth is left out entirely—a variation that is not only widely accepted, but supports the research that Western emoticons utilize the whole “face” while East Asian emoticons rely on the “eyes” to convey emotions

Democratizing Communication

As a whole, the cultural development of emoticons reinforces the need for a personal element in DMC. The nuances that emoticons add need to mean something to the audience—which is why a standard set of emoticons is not sufficient, even while the standard of using emoticons becomes widespread. The cultural development of emoticons also emphasizes how important emoticons have become to digitally-mediated communication and maintaining our social networks. The fact that we see cultural variations in emoticons reveals that emoticons are being used to connect people—to help people understand each other through methods that limit shared information inherent to social biofeedback.

As we increasingly transact our lives online, and find ways to effectively communicate online, what we are able to share becomes important in shaping our world overall. Researchers propose “the creation and distribution of digital goods has been democratized.”12 The creation and growth of social networking has allowed for easy sharing of creative and intellectual property. To which I will add, that these efforts have been assisted by DMC tools, such as emoticons, which allow shared ideas to be understood—not just in terms of content, but in terms of the author’s meaning as well.

Notes: 1. Wellman H, Liu. D. 2004. | 2. Roch-Levecq A.  2006. | 3. Wimmer H, Perner J. 1983. | 4. Fraiberg S. 1974. | 5. Minter ME, Hobson RP, Pring, L. 1991. | 6. Fogel A. 1977. | 7. Russell, Kay 2002. | 8. Wolf A. 2000. | 9. Kindred J, Roper S. 2004. | 10. Kim E, Gilbert S. 2009. | 11. Jack R, Blais C, Scheepers C, Schyns P, Caldara R. 2009. | 12. Bakshy E, Karrer B, and Adamic LA. 2009.

Bakshy E, Karrer B, and Adamic LA. Social influence and the diffusion of user-created content. Proceedings of the Tenth ACM Conference on Electronic Commerce (Stanford, California, USA, July 06 - 10, 2009). EC '09. ACM, New York, NY, 325-334.

Fogel A. (1977) Seeing and Being Seen.  In Blindness and Psychological Development in Young Children. Collis GM, Lewis W eds. Leicester: British Psychological Society.

Fraiberg S. (1974) Blind Infants and their Mothers: An Examination of the Sign System.  In The Effect of the Infant on Its Caregiver. M. Lewis and L. Rosenblum (eds).  New York: Wiley.

Jack RE, Blais C, Scheepers C, Schyns PG, & Caldara R (2009). Cultural confusions show that facial expressions are not universal. Current biology : CB, 19 (18), 1543-8 PMID: 19682907

Kim E, Gilbert S. Detecting Sadness in 140 Characters. The Web Ecology Project (2009) (accessed May 15, 2010)

Kindred J, Roper S. (2004). Making connections via instant messenger (IM): student use of IM to maintain personal relationships. Qualitative Research Reports in Communication, V, 48-54

Minter ME, Hobson RP, Pring, L. (1991). Recognition of vocally expressed emotion by congenitally blind children. Journal of Visual Impairment and Blindness, 85 (10), 411-415

Roch-Levecq A. (2006). Production of basic emotions by children with congenital blindness: evidence for the embodiment of theory of mind. British Journal of Developmental Psychology., 24, 507-528

Russell, Kay. Emoticons and internet shorthand. Computerworld 2002; 36(3).

Wellman HM, & Liu D (2004). Scaling of theory-of-mind tasks. Child development, 75 (2), 523-41 PMID: 15056204

Wimmer H, & Perner J (1983). Beliefs about beliefs: representation and constraining function of wrong beliefs in young children's understanding of deception. Cognition, 13 (1), 103-28 PMID:6681741

Wolf, A. (2000). Emotional Expression Online: Gender Differences in Emoticon Use CyberPsychology & Behavior, 3 (5), 827-833 DOI: 10.1089/10949310050191809

No comments:

Post a Comment