Exploring Revisions in Academic Text: Closing the Gap Between Process and Product Approaches in Digital Writing

Abstract

<jats:p> To date, research into dynamic descriptions of text has focused mainly on the spoken mode; and while writing process research has examined language structures, it has largely ignored the functionality (meaning) inherent in them. Therefore, drawing on systemic functional linguistics (SFL) and keystroke logging software, this article takes a further step toward an interdisciplinary dialogue by outlining a new schematic for coding and analyzing revisions. More specifically, we show how revision activity can be tracked within functional components, across functional components, and across clauses in terms of forward and backward movements. By exploring three digitally constructed texts, which were produced and observed unobtrusively in a natural setting, we have attempted to illustrate how one writer’s revising process can be operationalized in terms of (a) chronological movement (sequence) and (b) spatial movement (location). Findings showed how activity was relatively consistent across datasets with regard to session management, revision frequency, and distribution of revision types. Moreover, results also showed how most revision activity occurred at, or ahead of, the point of inscription, particularly with regard to revising the end of clauses. However, findings also indicated that revising the start of clauses was equally important when considering the size of functional components. </jats:p>

Digital Articulation: Examining Text-Based Linguistic Performances in Mobile Communication Through Keystroke-Logging Analysis

“…In addition to addressing the limitations discussed above, future study could continue to examine the data collected in different ways. For example, analysis of revision patterns could examine whether patterns found by Bowen and Van Waes (2020) may extend to texting, analysis of the timing processes of morpho-phonological features, such as -t/-d or -ing deletion, may examine Einstein’s (2015) observations from Twitter data, and more detailed analysis of pause bursts may yield greater insight into how the psychomotor processes of composing on mobile devices may differ from writing on computer-based keyboards ( Mangen et al, 2015 ; Galbraith and Baaijen, 2019 ). Such analyses would necessarily expand upon the tools used for this study, requiring means to parse keystroke data in order to identify revision bursts and morphological features from keybursts.…”

“…For example, participants in Leijten et al (2012) composed short simulated tweets; participants in Leijten et al (2015) composed texts typically written in under 10 min; participants in Van Waes et al (2010) revised short sentences; and participants in Nottbusch (2010) composed short sentences in response to stimuli. Even in studies that observed writers composing formal reports over multiple days, such as Leijten et al (2014) and Bowen and Van Waes (2020) , participants engaged in individual writing episodes that would last between 20 min to several hours. In other words, a study of writing on mobile devices, which may involve shorter forms of writing, may nevertheless yield similar data sets as Leijten et al (2014) and Bowen and Van Waes (2020) , in addition to involving writing more frequently throughout a participant’s everyday life.…”

“… Leijten et al (2014) additionally demonstrated that examining even a single writer producing a single text over a several-day period may yield important theoretical insights about how writers may use schematic knowledges of various genres and registers to construct texts. Most recently, Bowen and Van Waes (2020) used keystroke logging and systemic functional linguistics frameworks in order to examine revisions during writing, including the finding that revisions may most frequently occur at or just before the point of inscription. While Bowen and Van Waes (2020) also only observed a single writer composing over multiple writing sessions, their study demonstrated the rich possibilities of applying keystroke analysis in order to examine linguistic frameworks.…”

“…Most recently, Bowen and Van Waes (2020) used keystroke logging and systemic functional linguistics frameworks in order to examine revisions during writing, including the finding that revisions may most frequently occur at or just before the point of inscription. While Bowen and Van Waes (2020) also only observed a single writer composing over multiple writing sessions, their study demonstrated the rich possibilities of applying keystroke analysis in order to examine linguistic frameworks. Indeed, the amount of data obtained via keystroke logging from individual writers over longer spans of time, rather unobtrusively and indirectly, allows for in-depth analysis and consideration of how individuals may meaningfully construct written language in context in order to contribute to theory-building to conduct broader and more generalizable studies.…”

“…While only a short sample, these examples demonstrate both the value of using keystroke analysis to ask linguistic questions regarding text production. These studies may privilege lexical and morpho-syntactic analysis, partially due to the incorporations of NLP methods; however, as suggested by Bowen and Van Waes (2020) , Eisenstein (2015) , and Nottbusch (2010) , keystroke analysis remains promising for asking sociolinguistic questions. For example, how might the timing processes of enregistered and unmarked sociolinguistic variables differ ?…”