Table 21 below shows the comparison of feature frequency between synchronous, limited persistence media (chat and IM) and asynchronous, extended persistence media (forums, email, and SMS).
41 A Note about Predictive Texting as it Relates to the Interpretation of Tables 20 through 31: Predictive texting,
available on some mobile devices, “automatically anticipates later letters in a word, based upon the initial letters entered” (Ling & Baron, 2013, p. 203). Although predictive texting may exert an influence on typing speed and terminology used (i.e., that someone could intend to enter a shortened form, but the text might be expanded), the researcher cannot make assertions about its possible effects, because she did not have data on the types or capabilities of mobile devices used by the participants in the SMS section of the corpus, or if predictive texting increased or hampered texting speed.
Table 21: Comparison of features between synchronous and asynchronous media. Feature Synchronous / Limited Persistence (chat, IM) Asynchronous / Extended Persistence (forums, email, SMS) 𝝌2 Sig n % n % Acronyms / initialisms 1697 61.13% 1079 38.87% (1, N=2776) = 2.40, p = 0.121 Shortenings 1014 62.75% 602 37.25% (1, N=1616) = 0.03, p = 0.872 Clippings 163 62.93% 96 37.07% (1, N=259) = 0.02, p = 0.899 Single-‐letter forms 404 55.34% 326 44.66% (1, N=730) = 16.21, p = 0.000 ** Letter homophones 230 49.04% 239 50.96% (1, N=469) = 36.56, p = 0.000 ** Number homophones 43 42.16% 59 57.84% (1, N=102) = 18.12, p = 0.000 ** Symbolic Substitution 207 49.64% 210 50.36% (1, N=417) = 29.69, p = 0.000 ** Conjunctions / disjunctions 98 40.66% 143 59.34% (1, N=241) = 49.30, p = 0.000 ** Punctuation omission 691 67.75% 329 32.25% (1, N=1020) = 11.73, p = 0.001 * Non-‐standard use of
lowercase 1820 69.84% 786 30.16% (1, N=2606) = 59.04, p = 0.000 ** State abbreviations 45 88.24% 6 11.76% (1, N=51) = 14.36, p = 0.000 ** Onomatopoeic expression 509 75.07% 169 24.93% (1, N=678) = 45.37, p = 0.000 ** Phonetic respellings 562 69.21% 250 30.79% (1, N=812) = 15.37, p = 0.000 ** Offsetting punctuation 182 50.14% 181 49.86% (1, N=363) = 23.89, p = 0.000 ** All caps 984 80.07% 245 19.93% (1, N=1229) = 160.88, p = 0.000 ** Letter duplication 248 76.31% 77 23.69% (1, N=325) = 26.25, p = 0.000 ** Punctuation duplication 795 61.39% 500 38.61% (1, N=1295) = 0.75, p = 0.387 Spelling aloud 1 20.00% 4 80.00% -‐-‐ Emoticons 272 37.62% 451 62.38% (1, N=723) = 16.21, p = 0.000 ** Emotes 201 73.09% 74 26.91% (1, N=275) = 13.03, p = 0.000 ** Pointing 17 89.47% 2 10.53% (1, N=19) = 5.88, p = 0.015 Pictograms 0 0.00% 1 100.00% -‐-‐ Misspellings / typos 509 59.53% 346 40.47% (1, N=855) = 3.33, p = 0.068 Repairs 24 96.00% 1 4.00% (1, N=25) = 11.94, p = 0.001 * Addressivity 0 0.00% 5 100.00% -‐-‐ Reduplication 20 66.67% 10 33.33% (1, N=30) = 0.22, p = 0.642 Affixation / combining forms 93 65.03% 50 34.97% (1, N=143) = 0.38, p = 0.540 Compounds / space omission 491 47.62% 540 52.38% (1, N=1031) = 98.12, p = 0.000 ** Blends 0 0.00% 2 100.00% -‐-‐ Conversion 90 82.57% 19 17.43% (1, N=109) = 18.64, p = 0.000 ** Formatting workarounds 3 9.09% 30 90.91% (1, N=33) = 40.27, p = 0.000 ** Total 11413 62.55% 6832 37.45% ** p < .001 * p < .01
Media classified as being synchronous were also classified as having limited persistence, and media classified as being asynchronous were also classified as having extended persistence. Consequently, there are some similarities in these two sets of media characteristics. For example, face-to-face speech is synchronous and, typically, non- persistent (i.e., unless someone is recording the speech). The text communication in
synchronous, limited persistence media might then be thought of as more akin to speech than writing, as Hård af Segerstad (2002) and Davis and Brewer (1997) claim, for example. Thus, one might expect to find surrogate face-to-face cues in higher proportions in synchronous, limited persistence media than in asynchronous, extended persistence media, and the results in Table 21 support this. Onomatopoeic expression, phonetic respellings, all caps to convey shouting or emphasis, letter duplication, and emotes are some of the ways interlocutors may attempt to inject their own presence into the conversation and move it more toward a speech- like experience.
Furthermore, both synchronous and limited persistence media, like face-to-face speech, do not always allow interlocutors to look back on earlier utterances in the way interlocutors can when revising and editing a more persistent piece of writing. Thus, disfluencies, a natural part of speech often edited out in more traditional forms of writing, might also be expected to appear in higher proportions in synchronous, limited persistence media. However, differences in frequencies of misspellings and typos were not significant in these comparisons, which suggest that persistence and synchronicity have little influence over the production of errors; but the difference in frequency of repairs was significant. Repairs appeared in greater proportions in synchronous, limited persistence media. This
lends further support to Hård af Segerstad’s (2002) and Davis and Brewer’s (1997) assertions about the speech-like quality of synchronous, limited persistence media. Unplanned
discourse in face-to-face conversations may require more mid-conversation or in-situ repairs than in writing, which is often planned. Johnstone (2008) explains that there is a greater preponderance of repair mechanisms in relatively unplanned discourse as opposed to relatively planned discourse.
Herring (2002) explains that synchronous communication may lead to more phatic communication and these findings—specifically the higher proportions of onomatopoeic expression, phonetic respellings, all caps, letter duplication, and emotes—may support Herring’s assertion. Only one feature thought to convey emotion and sociality—emoticons— was found among those that are significant in asynchronous, extended persistence media. Additionally, emotes are performative, and allow interlocutors to communicate social and emotive information as Herring (2002) explains interlocutors are likely to do in synchronous settings. One of the motivations in using these face-to-face surrogates may be to convey a sense of one’s self and bridge the distance inherent in online communication.
Punctuation omission, non-standard use of lowercase, onomatopoeic expression, phonetic respellings, all caps, letter duplication, emotes, repairs, and conversion appeared in higher proportions in synchronous, limited persistence media. Synchronous media are thought to encourage interlocutors to be brief (Ferrara et al., 1991; Herring, 2002), and Herring (2007) claims that low visibility persistence may also increase the likelihood of abbreviations. Yet there are more high-frequency abbreviations (5 to be exact) with significant chi-square values in asynchronous, extended persistence media than in synchronous, limited persistence media (2, or 3 if phonetic respellings also reduce
keystrokes). Furthermore, many of the high proportion features in synchronous, limited persistence media—such as letter duplication, all caps, onomatopoeic expression, and emotes—do not save time or keystrokes, and therefore would not result in shorter messages. This suggests that synchronous, limited persistence media do not induce greater abbreviation and brevity than asynchronous, extended persistence media.
Asynchronous, extended persistence media would seem to allow interlocutors more time to plan, edit, and review messages before sending. Thus, Crystal (2006, p. 140) claims that extended persistence media may push language more toward that seen in “articles, books, and other ‘permanent’ literature.” “There is an autonomy about the text, once it is posted, much like that encountered in a book” (Crystal, 2006, p. 140). Because messages from extended persistence media may reside longer on interlocutors’ computers and may be shared with others, one might expect to see more adherence to standard rules of grammar and spelling because such persistent communication may act as record of one’s conduct.
Conversely, synchronous interactions are thought to cause the most “radical” linguistic innovations (Crystal, 2006, p. 135). When considering Crystal’s assertions, one would not expect to find many features in asynchronous, extended persistence media that bend the rules of typography and orthography. Yet in these data, several features that involve reshaping typography and orthography—single-letter forms, letter homophones, number homophones, offsetting punctuation, emoticons, compounds/space omission, and formatting
workarounds—appear in higher proportions in asynchronous, extended persistence media, and their chi-square values are significant. Perhaps as interlocutors have more time to read and re-read messages, they are more likely to play and experiment with typography and orthography. Only two features were more frequent in asynchronous, extended persistence
media that are also found in general/standard English texts with some degree of frequency: conjunctions/disjunctions and symbolic substitution. (However, common uses of punctuation and other non-alphabetic symbols were excluded from consideration during analysis of terms in the corpus and were also higher in proportion in extended media. See Appendix E: Signs and Symbols.) So although there is a possibility that one’s behavior can be tracked in extended persistence messages, this does not appear to be a motivation for avoiding typographical and orthographical deviations.
Furthermore, if one were to consider the use of a greater number of different kinds of cyberlanguage features to be evidence of greater innovation and less resemblance to
traditional writing, Table 21—which shows roughly equal numbers of significant features in both synchronous, limited persistence and asynchronous, extended persistence media— would not support Crystal’s assertion. Thus, probably the strongest link between
synchronicity and persistence centers on the speech-like qualities of synchronous media and limited persistence media.