Show Less
Open access

Conversational Writing

A Multidimensional Study of Synchronous and Supersynchronous Computer-Mediated Communication


Ewa Jonsson

The author analyses computer chat as a form of communication. While some forms of computer-mediated communication (CMC) deviate only marginally from traditional writing, computer chat is popularly considered to be written conversation and the most «oral» form of written CMC. This book systematically explores the varying degrees of conversationality («orality») in CMC, focusing in particular on a corpus of computer chat (synchronous and supersynchronous CMC) compiled by the author. The author employs Douglas Biber’s multidimensional methodology and situates the chats relative to a range of spoken and written genres on his dimensions of linguistic variation. The study fills a gap both in CMC linguistics as regards a systematic variationist approach to computer chat genres and in variationist linguistics as regards a description of conversational writing.
Show Summary Details
Open access


← 296 | 297 →


Appendix I.  Texts used in Biber’s (1988) study

Texts used in Biber’s (1988) study of language variation. The corpus totals approximately 960,000 words for 481 texts (Biber 1988: 67, 209–210, 1995: 87).

GenreTexts usedNumber of textsApprox. number of words
SpeechFace-to-face conv.texts 1.1–1.14 and 3.1–3.6 from LLC44115,000
Telephone conv.texts 7.1–7.3, 8.1–8.4 and 9.1–9.3 from LLC2732,000
Interviews124texts 5.1–5.3, 5.5–5.7, 6.1, 6.3, 6.4a, 6.5 and 6.6 from LLC2248,000
Broadcaststexts 10.1–10.7 and part of text 10.8 from LLC1838,000
Spont. speeches125texts 11.1–11.5 from LLC1626,000

← 297 | 298 →
Prepared speechestexts 12.1–12.6 from LLC1431,000
WritingPress reportageall texts in LOB category A4488,000
Press editorialsall texts in LOB category B2754,000
Press reviewsall texts in LOB category C1734,000
Religionall texts in LOB category D1734,000
Hobbiesthe first 30,000 words (texts 1–14) LOB category E1430,000
Popular lorethe first 30,000 words (texts 1–14) LOB category F1430,000
Biographiesthe first 30,000 words (texts 1–14) LOB category G1430,000
Official documentstexts 1–6, 13–14 and 25–30 from LOB category H1428,000
Academic proseall texts in LOB category J80160,000
General fictionall texts in LOB category K2958,000
Mystery fictionthe first 30,000 words (texts 1–14) LOB category L1326,000
Science fictionall texts in LOB category M612,000
Adventure fictionthe first 30,000 words (texts 1–14) LOB category N1326,000
Romantic fictionthe first 30,000 words (texts 1–14) LOB category P1326,000
Humorall texts in LOB category R918,000
Personal letterswritten to friends/relatives, collected by D. Biber66,000
Professional letterson administrative matters, collected by W. Grabe1010,000

← 298 | 299 →

Appendix II.  Descriptive statistics for genres studied

The frequencies in tables 1–7 are all normalized to text lengths of 1,000 tokens (except for type/token-ratio and word length); see section 3.2.

Table 1:  Descriptive statistics for Internet relay chat


← 299 | 300 →


← 300 | 301 →

Table 2:  Descriptive statistics for split-window ICQ chat


← 301 | 302 →


← 302 | 303 →

Table 3:  Descriptive statistics for the SBC subset (spoken American English)


← 303 | 304 →


← 304 | 305 →

Table 4:  Descriptive statistics for Biber’s corpus as a whole (Biber 1988: 77–78)


← 305 | 306 →


← 306 | 307 →

Table 5:  Normalized frequencies per Internet relay chat text


← 307 | 308 →


← 308 | 309 →

Table 6:  Normalized frequencies per split-window ICQ chat text


← 309 | 310 →


← 310 | 311 →

Table 7:  Normalized frequencies per text in the SBC subset (spoken Am. English)


← 311 | 312 →


← 312 | 313 →

Appendix III.  Raw frequencies of linguistic features

Tables 1a–3a present the raw frequencies per text of the linguistic features in the corpora investigated (for type/token ratio and word length, see Appendix II). The length of each text is shown in tables 1b–3b.

Table 1a:  Raw frequencies per Internet relay chat text


← 313 | 314 →


Table 1b:  Length of the Internet relay chat texts


← 314 | 315 →

Table 2a:  Raw frequencies per split-window ICQ chat text


← 315 | 316 →


Table 2b:  Length of the split-window ICQ chat texts


← 316 | 317 →

Table 3a:  Raw frequencies per text in the SBC subset (spoken American English)


← 317 | 318 →


Table 3b:  Length of the SBC subset texts (spoken American English)


← 318 | 319 →

Appendix IV.  Examples of excluded material

Certain messages and strings of text were excluded from the conversational writing logs and the SBC subset before the texts were annotated for the features in Biber’s (1988) methodology. Typical excluded instances are exemplified below.

Excluded from Internet relay chatExample
Bracketed nickname turn indicators<River>
Server-generated messagesSession Start: Mon Mar 25 18:01:47 2002
(session start messages, time stamps, join- and leave messages)[18:01] *** Now talking in #family
*** edi-tr has joined #family

*** edi-tr has quit IRC (Killed (NickServ (Nickname Enforcement)))
Channel operator interference*** ezococx was kicked by Sp0ck (banned: spam)
Action commands* big-dog º©o¿,,¿o©º°¨¨°º© HELLO WELCOME TO>>>
(including graphic noise) #family CHANNELº<<< memyselfandi ©º°¨¨°º©o¿,,¿o©
* NA_TuPaC slaps ma7ash around a bit with a large trout
* SwampRocker Bunny.. (Y) … (Y).. ^gypsy^ . (Y)….(Y).. Hugs

* SwampRocker Bunny . (°.°) . (°.°). ^gypsy^ .(°.°).. (°.°). Hugs

* SwampRocker Bunny .()¯°¯() ()¯°¯() ^gypsy^ ()¯°¯() ()¯°¯() Hugs

* SwampRocker Bunny .(_)-(_) (_)-(_) ^gypsy^ (_)-(_) (_)-(_) Hugs
Foreign languagealguien habla espaÑol???
Excluded from split-window ICQExample
Bracketed nickname turn indicators<5>
Action tropes2 points finger at you, scolding you for your actions B picks a flower and hands it to you
Foreign languagehablamos espanol abren los libros a la paginaaaaaaaaaaaaa tres
Excluded from face-to-face SBCExample
Foreign languageQue es mas o menos. No es exelente, pero es mas o menos. ← 319 | 320 →

Appendix V.  Features with a |standard score| >2.0

Table 1 lists the features with a standard score above 2.0, or below -2.0, in the genres studied, the most influential (most salient) contributors to the dimension scores of the particular genre. Section 4.4, and part of 4.2, explore the most salient features of the conversational writing genres (split-window ICQ chat and Internet relay chat) and present their distribution in writing, ACMC and speech. The procedure of standard score calculation is described in section 3.5.

Table 1:  Features with a |standard score| >2.0 in the genres studied. A hyphen (-) indicates that the genre has no feature with a |standard score| >2.0


← 320 | 321 →


← 321 | 322 →

Appendix VI.  Statistical tests of salient features

Table 1 presents the values for probability (p) from t-tests of the feature distributions in SCMC, SSCMS, writing and speech for the salient features in conversational writing discussed in chapter 4. For some of the features, or combinations of features, p-values are not available (“n.a.”) owing to the unavailability in Biber (1988) of the requisite data for the test. As regards inserts, no annotation of Biber’s (1988) texts of writing or speech was carried out; instead, the p-values for inserts given in table 1 in the comparisons to “speech” reflect for “speech” only the face-to-face conversations from SBC (as noted in section 4.6). With regard to emotives, the tests here reflect that none of the written (LOB) or spoken (LLC or SBC) texts contains emotives.

Table 1:  Values for probability (p) from t-tests of features in writing, speech, SCMC and SSCMC. Significant differences in bold (p<.05). The p-values have not been multiplicity adjusted


← 322 | 323 →

Appendix VII.  Word lists for the corpora studied

Table 1:  Word frequency lists for the corpora studied: IRC, split-window ICQ chat and the SBC subset, by rank for the fifty most frequent types (not lemmatized). N.B. raw frequencies for full corpora


← 323 | 324 →


← 324 | 325 →

Appendix VIII.  Dimension score statistics for Biber’s (1988) genres

Dimension 1:Informational versus Involved Production
Dimension 2:Narrative versus Non-Narrative Concerns
Dimension 3:Explicit/Elaborated versus Situation-Dependent Reference
Dimension 4:Overt Expression of Persuasion/Argumentation
Dimension 5:Abstract/Impersonal versus Non-Abstract/Non-Impersonal Information
Dimension 6:On-Line Informational Elaboration

Table 1:  Descriptive dimension statistics for Biber’s genres (Biber 1988: 122–125)


← 325 | 326 →


← 326 | 327 →


← 327 | 328 →


← 328 | 329 →


← 329 | 330 →

Appendix IX.  Computation of cluster affiliations

Table 1 shows the centroid scores of each cluster identified in Biber (1989, 1995) with respect to Biber’s (1988) Dimensions 1 through 5 (C1 means cluster centroid 1, C2 cluster centroid 2, etc.). Tables 2–4 each present the Euclidean distances found between the texts and the cluster centroids (and those between the average dimension scores of the genre and the latter), with the resulting cluster affiliations indicated in the rightmost column. Table 5 presents the Euclidean distances found between the dimension scores of Collot’s (1991) genre of BBS conferencing (i.e. the “ELC other” corpus of ACMC) and the cluster centroids. The polarity of all scores follows that in Biber (1988, 1989), rather than that in Biber (1995). See Appendix X for the dimension scores of the individual texts, and tables 5.1 and 5.5 (in chapter 5) for those of the genres.

Table 1:  Cluster centroid scores (Biber 1995: 328–331)126


Table 2:  Distances to cluster centroids of the IRC texts (and the IRC genre)


← 330 | 331 →

Table 3:  Distances to cluster centroids of the split-window ICQ texts (and the splitwindow ICQ genre)


Table 4:  Distances to cluster centroids of the SBC subset texts (and the SBC subset genre)


Table 5:  Distances to cluster centroids of the BBS conferencing genre (“ELC other” corpus, Collot 1991)


← 331 | 332 →

Appendix X.  Dimension scores for individual texts

Tables 1–3 present the dimension scores on Biber’s (1988) dimensions for the individual texts annotated in the present study.

Table 1:  Dimension scores for the Internet relay chat texts (UCOW)


Table 2:  Dimension scores for the split-window ICQ texts (UCOW)


Table 3:  Dimension scores for the SBC subset texts annotated in the present study


124 “Interviews” denotes public conversations, debates and interviews (Biber 1988, 1995: 87).

125 In his description of the sampling procedure, Biber (1988: 210) indicates that spontaneous speeches were divided into 15 texts, which would yield a total of 480 texts. Later accounts, however, maintain that there was a total of 481 texts (e.g. Biber 1995: 87, Conrad & Biber 2001: 111), which explains why the figure from Biber (1988: 67) is retained here.

126 The polarity of all scores follows that in Biber (1988, 1989).