ArticlePDF Available

Abstract

This review discusses and compares the findings of 38 peer-reviewed studies on text-based synchronous computer-mediated communication for second or foreign language learning from over the past twelve years. Research themes that emerged include: modality, corrective feedback, noticing, alignment and uptake, as well as task design/conditions and grouping. The comparison of findings revealed that the medium may enhance or hinder language learning depending on the context, method of implementation , and pedagogical aims. It also uncovers implications for language learning theory and practice that are less evident when examining each study in isolation, including findings on face-preservation and learner engagement, features of communication besides negotiation of meaning that lead to language acquisition, as well as a possible interaction between cognitive complexity and modality in the effectiveness of language learning tasks.
Technology in Language Teaching & Learning
ISSN 2652-1687
https://www.castledown.com/journals/tltl/
Technology in Language Teaching & Learning, 4(1), 1–17 (2022)
https://doi.org/10.29140/tltl.v4n1.712
Copyright: © 2022 Hughes. This is an open access article distributed under the terms of the Creative
Commons Attribution Non-Commercial 4.0 International License, which permits unrestricted use,
distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability Statement: All relevant data are within this paper.
Text-Based SCMC for SLA:
A Narrative Review
LEANDER S. HUGHES
Saitama University, JAPAN
hughes@mail.saitama-u.ac.jp
Abstract
This review discusses and compares the ndings of 38 peer-reviewed studies on text-based synchronous
computer-mediated communication for second or foreign language learning from over the past twelve
years. Research themes that emerged include: modality, corrective feedback, noticing, alignment and
uptake, as well as task design/conditions and grouping. The comparison of ndings revealed that the
medium may enhance or hinder language learning depending on the context, method of implementa-
tion, and pedagogical aims. It also uncovers implications for language learning theory and practice that
are less evident when examining each study in isolation, including ndings on face-preservation and
learner engagement, features of communication besides negotiation of meaning that lead to language
acquisition, as well as a possible interaction between cognitive complexity and modality in the eec-
tiveness of language learning tasks.
Keywords: Text-based synchronous computer-mediated communication (SCMC); second language
acquisition (SLA); interaction; modality; aordances
Introduction
With the onset of the COVID-19 pandemic, computer-mediated communication (CMC) and partic-
ularly synchronous CMC (SCMC) has at times become the primary medium of education in many
regions around the world, with video-based SCMC platforms such as Zoom and Google Meet taking
center stage. While video-based SCMC has a strong corollary in traditional face-to-face communication
(FTFC), text-based SCMC (TSCMC) is unique in that it transforms writing, a traditionally asynchro-
nous form of communication, into a synchronous medium. It thus combines the real-time interactivity
of spoken language with the permanent and textual nature of writing. Due to this unique combination of
Hughes: Text-Based SCMC for SLA: A Narrative Review 2
properties TSCMC oers novel aordances for language learning and research, and the body of studies
exploring the potential of TSCMC for second language acquisition (SLA) continues to grow.
Prior reviews of TSCMC research
Several research reviews have already been conducted that investigate CMC for SLA (Lin, 2014; Lin,
2015; Lin, Huang, & Liou, 2013; Sauro, 2011; Zeigler, 2016). Of these, only two (Sauro, 2011; Lin
et al., 2013) allow the reader to distinguish the ndings for TSCMC from other forms of CMC.
Sauro’s research synthesis
Sauro (2011) reviewed 97 CMC studies from peer-reviewed journals. Notable ndings for TSCMC
include the observation that learners used more discourse functions in TSCMC compared to other
modalities and that interaction via TSCMC led to greater gains in learners’ ability to perform the
refusal speech act compared to spoken modalities. In addition to these were results that showed
learners who engaged in TSCMC achieved greater gains in uency than those in a control or FTFC
group and that negotiation in TSCMC decreased as scrolling and cursor movement increased, as
well as the observation that learners produced more complex output in TSCMC after self- initiated
deletions compared to those that were partner-initiated through interruption. Sauro concludes by
pointing out future avenues of research into SCMC, recommending studies involving younger
participants, learners of non-alphabetic languages, and pedagogic tools that are unique to SCMC
contexts.
Sauro (2011) provides interesting details from TSCMC studies and attempts to respond to seminal
questions for CALL research posed by Chappelle (1997) including, “How good is the language expe-
rience in CALL for L2 learning?” (p. 28). However, no overall ndings with regard to TSCMC or
SCMC in general are provided in response to this question. The review concludes by giving counts
of studies investigating each of the four components of Canale and Swain’s (1980; Canale, 1983)
communicative competence without indicating how eective SCMC might be for acquiring them. The
review’s inability to provide a satisfactory answer to Chappelle’s question may have been due in part
to the sheer breadth of studies included, making it dicult to investigate each study in enough depth
to form conclusions about the eectiveness of the technology.
Lin and colleagues’ meta-analysis
Lin et al. (2013) conducted the only meta-analysis to date exclusively on the eect of TSCMC on
SLA. It included 13 studies and found a modest overall eect size (m = .33) for conditions or interven-
tions employing TSCMC. It also indicated two factors that appeared to moderate the eectiveness of
TSCMC. First, only interventions that lasted a week or longer produced any eect. Second, TSCMC
was more eective when done in pairs as opposed to larger groups. However, the researchers caution
that these ndings should be considered tentative, due to the small sample size of studies included
in the meta-analysis. Despite this weakness, Lin and colleagues do manage to provide an answer to
Chappelle’s question regarding the use of the technology for L2 learning: TSCMC has a signicant,
yet weak eect on SLA.
This conclusion, however, oversimplies the case for TSCMC. An inspection of the eect sizes from
each individual study included reveals that they range from strongly positive to strongly negative
indicating that TSCMC can be very eective in some cases while ineective or even harmful in oth-
ers. Beyond the tentative support for pair work and longer interventions, we are left wondering what
contextual factors may account for the enhancing or hindering eects of TSCMC.
3 Technology in Language Teaching & Learning, 4(1)
The Present Study
Nearly a decade has passed since the meta-analysis by Lin et al. (2013) and while many more studies
on TSCMC have emerged, no review has since attempted to investigate for overall patterns or trends
between studies in order to answer the question of how TSCMC may enhance or hinder language
learning. Furthermore, despite the mode’s uniqueness, not only as a means of interaction, but also
as a tool for testing language learning hypotheses, no review has yet investigated the contribution of
TSCMC studies to SLA theory.
This study therefore reviews and compares the ndings of research on TSCMC from over the past
twelve years in order to answer the following questions.
1. What aspects of SLA appear to be enhanced or hindered by TSCMC?
2. What are the insights provided and questions raised with regard to SLA theory and
practice?
Method
While the rigor of the reviews by Sauro (2011) and Lin et al. (2013) is admirable, their adherence
to synthetic approaches may have inadvertently limited the meaningfulness of their ndings. Norris
and Ortega (2007) present a generally favorable account of research synthesis, but warn of its pit-
falls which may produce “technically sound but theoretically impoverished research” (p. 810). In
order to tease out hitherto unseen patterns or connections, the current review prioritizes exibility
and interpretation over systematicity and rigor. It therefore takes a narrative, rather than synthetic
approach. Even so, this review does employ a number of techniques common to research syntheses,
including a systematic retrieval method and inclusion criteria. These are all utilized in the service
of the narrative however, and as such, are applied more exibly than they would be in a research
synthesis.
Retrieval of studies
Following Sauro’s (2011) approach, 16 peer-reviewed journals which Smith and Laord (2009) and
their informants had identied as the highest ranked journals featuring CALL and CALL related con-
tent were chosen as the primary sources for this review.
These journals were searched for the keywords: synchronous, computer, mediated, communication,
text, and chat with “computer-mediated communication” quoted exactly when possible. The year 2010
was chosen as the starting point for articles to include and studies that were covered in the reviews by
Sauro (2011) or Lin et al. (2013) were excluded from this one to avoid overlap.
In addition to the above journal searches, Google Scholar was searched using the same terms and
the rst 100 results examined for further publications to include, with an additional criterium that
publications from this search had to have been cited by at least 20 other papers to qualify. This search
produced three further publications, each from a journal that had not been included in Sauro’s (2011)
review.
Inclusion criteria
Due to the focus on studies with direct implications for SLA, this review adopts three selection criteria
from Smith’s (2017) outline of CALL studies that are exemplary in their attempt to develop and extend
Hughes: Text-Based SCMC for SLA: A Narrative Review 4
both theory and practice in the eld. Specically, Criteria 6 through 8 are drawn from Smith’s paper,
with Criterium 7 undergoing some modication for clarity. The inclusion criteria for this review are
thus that each study had to:
[Stage 1]
1. have been published in one of the 16 peer-reviewed journals included in Sauro’s
(2011) review or be one of the three additional papers discovered through the Google
Scholar search explained above.
2. have been published in 2010 or later and have not been included in the previous
reviews by Sauro (2011) and Lin et al. (2013).
[Stage 2]
3. be a primary study, not a research review or theoretical position paper.
4. involve at least one language learning condition during which the sole mode of
communication was TSCMC and include an analysis of the results of that condition
as distinct from those of any other condition
5. have collected data directly from the TSCMC and/or pre-post measures.
[Stage 3]
6. be “rmly grounded in an identiable SLA theoretical framework”
(Smith, 2017, p. 447).
7. “examine key assumptions, constructs, hypotheses, and so on which are relevant to
this theory of SLA” in that they are argued based on that theory to either lead to or
moderate the learning of language (Smith, 2017, p. 447).
8. “investigate some aspect of the argued aordances of CALL environments”
(Smith, 2017, p. 447).
In Stage 1 of the selection process, the initial search described above produced over 327 titles which
satised Criteria 1 and 2. In Stage 2, a reading of the abstracts and papers revealed that 53 satised
Criteria 3 through 5. In Stage 3, these papers were read in detail. Thirty-eight of them satised Criteria
6 through 8 and were thus included for review (see Appendix).
It is important to note that, although this method of selection parallels that used by Sauro’s (2011) research
synthesis, it was not meant to produce an exhaustive body of research on TSCMC but only a substantial
slice. Providing enough detail about each study to elucidate common or contrasting ndings was a prior-
ity of this narrative review and would not be possible if the number of studies included became too large.
Categorization
Papers were grouped based on similarities between their research questions, foci, and aims into ve
thematic categories:
Modality (k = 4)
Corrective Feedback (k = 3)
Uptake, Noticing, and Alignment (k = 9)
5 Technology in Language Teaching & Learning, 4(1)
Task Design and Conditions (k = 14)
Grouping (k = 9)
These themes do not cleanly divide the papers, however. For example, there were two studies that
investigated both the eects of modality and corrective feedback. These were assigned to the lower
Corrective Feedback section. Thus, categories lower on the list may contain studies that partly involve
categories higher on the list. The reverse however is not the case, with one exception: the study by
Dao, Duong, and Nguyen (2021) which is included in two dierent sections, because it was deemed
essential to the conclusions of both.
Results
Below is a review of the research within each of the ve themes that emerged. Note that because
30 studies involved university learners working in pairs, ages and interactive group sizes will not be
given unless other ages, backgrounds, or group sizes were involved.
Modality (k = 4)
Four studies investigated the eect of modality as their primary focus. Kim (2012) found that
more communication strategies (CSs) were used during communicative tasks in FTFC than in
TSCMC except for avoidance strategies which were more frequent in TSCMC. TSCMC was
thought to elicit more avoidance strategies because further negotiation of meaning (NoM) might
have disrupted the ow of conversation which, due to the nature of TSCMC, was already subject
to time-delays and split turns. Similarly, Hung and Higgins (2015) found that open conversation
through video-based SCMC led to a signicantly higher number of CSs compared to TSCMC.
They note, however, that communication problems in TSCMC were often resolved through self-
repair rather than CSs.
Ajabshir (2019) conducted a study wherein learners watched video clips on the request speech act and
then collaboratively answered questions about and composed dialogues based on the clips through
TSCMC, asynchronous CMC (email) or FTFC. The two CMC groups both achieved signicant pre-
post gains in the ability to perform the speech act over the FTFC group. The researcher posits that
CMC enhanced learning by reducing learners’ dependence on the teacher, but acknowledges that the
novelty of learning though CMC may have also played a role. Also, group assignment was based on
learner preference, which may have confounded the results.
Finally, Kessler, Polio, Xu, and Hao (2020) heeded Sauro’s (2011) call for studies involving learners of
non- alphabetic languages. Learners of Chinese discussed topics for 20 minutes in preparation to write
about them through TSCMC and FTFC. FTFC produced over three times more output than TSCMC,
while post-FTFC writing featured signicantly higher complexity. The disadvantage of TSCMC
stemmed primarily from the delay caused by typing non-alphabetic characters.
To summarize, Kim (2012) and Hung and Higgins (2015) suggest that CS use in TSCMC is less
prevalent than in other modalities, while Ajabshir (2019) found that TSCMC may enhance language
development when used to collaboratively synthesize and apply previously studied content. Finally,
Kessler et al. (2020) indicate that increased typing delays during TSCMC in non-alphabetic languages
may dramatically decrease its ecacy.
Hughes: Text-Based SCMC for SLA: A Narrative Review 6
Corrective feedback (k = 3)
Three studies involving TSCMC focused primarily on corrective feedback (CF). Darhower (2014)
provided CF in the form of dynamic assessment (DA) to learners of Spanish collaborating via
TSCMC to retell clips from a lm. Learner’s ability to narrate past events appeared to improve as
the amount of DA mediation required to resolve linguistic problems decreased over the six-week
study period.
Bower and Kawaguchi (2011) conducted a tandem study involving learners of English and Japanese
conversing through TSCMC. Chat scripts revealed that, whereas 7.2% to 20% of the turns were for
the negotiation of meaning (NoM), less than 5% of the linguistic errors made received CF. After the
chat sessions, learners corrected their partner’s chat script resulting in over 60% of the errors receiv-
ing CF. The researchers conclude that such post-chat correction might be considered to add focus on
form (FoF).
Arroyo and Yilmaz (2018) investigated immediate versus delayed CF during an information gap task
through TSCMC. Their immediate CF group achieved signicantly higher pre-post gains than the
delayed CF group whose gains were in turn signicantly higher than a no-CF control group.
These studies suggest that the interactive, yet textual and permanent nature of TSCMC may enhance
CF. While CF occurs rarely in learner-learner communication, TSCMC lends itself well to the provi-
sion of CF by an instructor or knowledgeable peer. While immediate CF was superior, delayed CF still
resulted in signicant gains, and thus might be considered when immediate CF is infeasible.
Uptake, noticing, and alignment (k = 9)
Ten studies featured a prominent focus on uptake, noticing, or alignment. Gigure and Parks (2018)
focused on younger learners as per Sauro’s (2011) recommendation. Their study involved Grade 6
elementary school learners who engaged in a tandem information exchange via TSCMC one hour per
week for six weeks. Despite the researchers actively encouraging learners to provide and respond to CF,
it occurred only rarely during chat sessions, and was seldom followed by a uptake in the form of repair.
Rouhshad, Wigglesworth, and Storch (2016) compared interaction in TSCMC to FTFC during an
information gap and decision-making task. TSCMC produced approximately half the number of words
as FTFC in over twice the time. Even after controlling for output amount, FTFC resulted in signi-
cantly more NoM and a higher percentage of uptake. However, uptake seldom occurred in either con-
dition, possibly due to the type of task used which emphasized meaning over form. To encourage more
focus on form (FoF), the researchers recommend collaborative writing tasks.
Following that recommendation, Dao, Duong et al. (2021) had 100 EFL learners aged 15 to 33 under-
take a picture sequencing and collaborative story writing task in TSCMC versus video-based SCMC.
Signicant dierences emerged in favor of FTFC for both CF and uptake. Video interaction produced
between two and three times the instances of CF found in TSCMC during the 30-minute timeframe.
Unlike Rouhshad et al. (2016) however, the percentage of CF instances followed by an accurate repair
was nearly equal for the two modalities.
Yuksel and Inan (2014) had 64 learners engage in a similar task and investigated whether TSCMC
enhanced noticing compared to FTFC. Although participants took more time in TSCMC and produced
fewer NoMs, they were able to accurately identify signicantly more instances of NoM in their com-
munication afterward during a stimulated recall.
7 Technology in Language Teaching & Learning, 4(1)
Smith (2012), investigated the relationship between noticing CF during TSCMC and language learn-
ing. Beforehand, Smith trained participants to recognize and learn from recasts. Learners then retold
a story for the researcher who provided frequent CF (recasts). Both eye-tracking and stimulated recall
data predicted immediate and delayed post-test scores, evidencing the relationship between CF, notic-
ing, and language acquisition.
Kim, Jung, and Skalicky (2019) and Kim, Skalicky, and Jung (2020) paired with learners and attempted
to induce alignment to grammatical structures during information gap tasks through TSCMC versus
FTFC. Alignment occurs when learners adjust their language to match their interlocutor’s due, osten-
sibly, to linguistic priming by a prior message. Kim et al. (2019) found that alignment for stranded
prepositions in relative clauses was signicantly more likely to occur during TSCMC compared to
FTFC. A marginally signicant dierence also emerged on an immediate post-test in favor of the
TSCMC group. A similar study by Kim et al. (2020) targeted direct and indirect questions, and found
a signicant dierence in alignment between groups in favor of TSCMC, but only for direct questions.
While the amount of alignment signicantly predicted immediate and delayed post-test scores for
both question types, modality itself did not signicantly predict scores for either type of question. The
results of Kim et al. (2019) and Kim et al. (2020) suggest that alignment does indeed lead to language
acquisition and that TSCMC can foster more alignment than spoken interaction, but that this eect
varies depending on the linguistic items targeted.
Uzum (2010) investigated alignment during TSCMC tasks between nine intermediate ESL learn-
ers and advanced or native English-speaking students. The researcher indicated that alignment
occurred frequently, but the construct was operationalized very broadly including, for example,
alignment in tone.
Kourtali (2022) investigated the eect of TSCMC on uptake during two information gap tasks done
with the researcher and subsequent language gains. The study is one of the few that involves younger
learners (aged 10 to 13). Despite the researcher providing more recasts during TSCMC, FTFC produced
signicantly more uptake and higher written and oral pre-post gains. Note though that the learners in
this study had never used TSCMC for language learning purposes before the experiment and received
no prior instruction or training regarding recognizing and responding to recasts in that medium.
In summary, Gigure and Parks (2018), Rouhshad et al. (2016), and Yuksel and Inan (2014) indicate
that FTFC elicits more NoM, CF, and uptake than TSCMC. However, this dierence may disappear
when output amount is controlled for (Dao, Duong et al., 2021), suggesting that it is likely caused by
the delay necessitated by reading and typing in TSCMC. Furthermore, Yuksel and Inan (2014) indi-
cate that TSCMC may be more eective for promoting noticing than FTFC. Noticing during TSCMC
was, in turn, shown by Smith (2012) to lead to language acquisition, as was alignment by Kim and
colleagues (2019, 2020) who also found that TSCMC leads to more alignment and, depending on the
target form, more acquisition than FTFC. Finally, the low uptake and linguistic gains for Kourtali’s
(2022) TSCMC group may have been due to their lack of experience with the modality, implying the
need for prior orientation and training.
Task design and conditions (k = 14)
The 14 studies on task design and conditions investigated planning, task type/characteristics, environ-
ment, and interaction.
Planning (k = 3). Hsu (2012, 2015) and Zeigler (2018) employed picture narration tasks through
TSCMC and investigated for an eect of pre-task planning on complexity, accuracy, and uency (CAF)
Hughes: Text-Based SCMC for SLA: A Narrative Review 8
of output. Hsu (2012, 2015) involved adult learners in an intensive English program in the U.S. aged
18 to 55. Hsu (2012) had learners take notes on paper for 10 minutes which were taken away before
learners paired with the researcher for the task. No signicant eects emerged.
Hsu (2015) had learners plan by rehearsing the entire task for 10 minutes in Microsoft Word which
they closed before doing the task with the researcher. This planning method produced a signicant
increase in grammatical verb accuracy during the task and clause length during a subsequent task.
Zeigler (2018) had participants take notes for 3 minutes, 1 minute, and not at all. Learners were paired
with each other and allowed to plan collaboratively, taking notes in the same chat window where they
remained accessible throughout the task. Planning for 3 minutes produced a signicant increase in
lexical complexity but did not inuence accuracy or uency.
To summarize, Hsu (2012) found that planning by taking notes for 10 minutes that were then taken
away had no eect. However, rehearsing the entire task on the computer for 10 minutes did produce a
signicant eect (Hsu, 2015). Zeigler (2018) also obtained signicant results for just 3 minutes of plan-
ning. Hsu’s (2015) rehearsal condition may have reduced cognitive load from meaning focus (Skehan,
2009) and allowed the practice of verb forms for increased grammaticity, whereas the shared nature and
short duration of the 3-minute planning in Zeigler’s (2018) study likely led to an emphasis on meaning,
producing a brainstorming eect which increased lexical complexity but otherwise did not aect CAF.
Task Type/Characteristics (k = 5). Yilmaz (2011) and Zeng (2017) compared dictogloss versus jig-
saw tasks on their facilitation of language related episodes (LREs) where learners discussed or cor-
rected the language they used. The dictogloss task engendered signicantly more LREs, and, in the
case of Yilmaz, correctly solved LREs. Zeng also found that TSCMC resulted in a higher frequency of
LREs than FTFC (though taking nearly three times longer). Learners commented that TSCMC allowed
them to look back on, notice, and address errors that might have been missed during FTFC.
Kim (2017) investigated the eect of task type and modality on question production and article use,
comparing spot-the-dierence, story retelling based on a picture sequence, and decision-making tasks
through TSCMC versus FTFC. A signicant dierence emerged in accurate article use in favor of the
spot-the-dierence followed closely by story retelling. Kim suggests that the dierence stemmed from
the need for more descriptive accuracy in those tasks, and their lower demand for logical reasoning
and negotiation allowing for more attention to accuracy. Modality also had an eect: TSCMC resulted
in a signicant increase in advanced question production and accurate article use. In FTFC, learners
could often get by with one-word questions, whereas overlapping turns and the absence of paralin-
guistic cues during TSCMC necessitated the provision of more context and specicity and thus more
advanced question forms in order to avoid confusion. This need for specicity combined with the extra
time between messages may have also increased attention to article accuracy. Importantly, and unlike
other studies, the amount of time taken to complete the tasks was nearly equal for the two modalities.
Bandl (2012) compared the accuracy and amount of output elicited by two dierent versions of a fam-
ily tree creation task through TSCMC versus asynchronous CMC. One version of the task was a jig-
saw information gap where pairs exchanged information to complete a ctional family tree. The other
version was a free information exchange about learners’ own family trees. Because the jigsaw task
required information to be exchanged for completion, Bandl hypothesized it would elicit more output.
However, the information exchange task produced signicantly more output, ostensibly because it was
more interesting for learners to talk about their own family trees than about a ctional one. A signi-
cant dierence (marginal in the case of TSCMC) in accuracy emerged in favor of the jigsaw task, but
this may have been because the jigsaw worksheets included additional target language input.
9 Technology in Language Teaching & Learning, 4(1)
Finally, Baralt (2013) paired with learners for a story retell task, with some doing a more cognitively
complex version which required inferring rather than simply reading the intentions of one of the charac-
ters. Baralt hypothesized based on Robinson’s (2001, 2011) Cognition Hypothesis that cognitive com-
plexity would increase language development by demanding more syntactically complex output and
thereby increasing learners’ attention and responsiveness to input and feedback. Indeed, the more cog-
nitively complex task when done through FTFC led to signicantly more language development than
two of the other experimental conditions. However, when carried out through TSCMC, the cognitively
simpler version led to more language development and, in fact, resulted in the highest language gains
of any condition. These ndings suggest that the eect of cognitive complexity may dier depending
on modality and that a task cognitively optimized for FTFC might be suboptimal for TSCMC and vice
versa. Moreover, the results complicate our understanding of the role of cognitive complexity in lan-
guage learning with those for FTFC seemingly supporting Robinson’s cognition hypothesis, whereas
those for TSCMC support Skehan’s competing Trade-O Hypothesis (1998, 2009).
In summary, Yilmaz (2011) and Zeng (2017) found that dictogloss, where initial target input is pro-
vided and reasoning requirements limited, led to an increased number of LREs and, in the case of
Yilmaz, correctly solved LREs. Zeng also found that TSCMC led to more frequent LREs than FTFC.
Kim (2017) discovered that doing a spot-the-dierence task where context is fully provided, greater
specicity required, and logical reasoning minimal led to more accurate article use than other tasks
and that TSCMC outperformed FTFC in terms of both article accuracy and question formation and
took nearly an equal amount of time. Bandl (2012) attempted to investigate the eect of interaction
being required versus optional to complete a task, but other factors may have confounded the results.
Finally, Baralt (2013) obtained evidence that, contrary to ndings for FTFC, decreasing the cognitive
complexity of tasks done through TSCMC may increase their eectiveness for language acquisition.
This study, in particular, is worthy of replication and extension as it carries important implications for
theories of cognition in SLA.
Environment (k = 3). The studies focusing on environment dealt with virtual worlds. Peterson (2012)
investigated TSCMC between learners during communicative tasks in the 3D multiuser virtual envi-
ronment (MUVE) Second Life for collaborative interaction hypothesized by sociocultural theory to
foster language development (Ohta, 1995). Examples observed included lexical and corrective peer
scaolding, the use of continuers prompting further elaboration, promotion of social cohesion and the
establishment of intersubjectivity. Peterson further observed that learners actively utilized their avatars
within the MUVE which appeared to enhance their engagement during the tasks.
Rama, Black, Van Es, and Warschauer (2012) investigated aordances for language learning in the
massively multiplayer online game (MMOG) World of Warcraft, identifying three main aordances: a
safe space for learning, an emphasis on communicative competence, and goal-oriented collaboration.
The study also contrasted the experience of one participant who was a beginner at the target language
but an experienced gamer and another with the opposite characteristics. The MMOG appeared highly
benecial for the gamer, but less so for the other participant.
Collentine (2013) investigated the eect of input complexity in a custom 3D mystery game on learner
output produced via TSCMC during their attempt to solve the mystery following game play. Lin-
guistically simple and complex versions of target language messages were created for characters and
descriptions in the game and randomly chosen for players such that complexity varied from learner to
learner. Collentine hypothesized that linguistically complex input during the game would promote lin-
guistically complex output. However, regression analyses showed that a large amount of linguistically
simple input optimally facilitated output during subsequent interaction. Collentine cautions, though,
that these results may have been confounded by initial prociency dierences between learners.
Hughes: Text-Based SCMC for SLA: A Narrative Review 10
To conclude, Peterson (2012) and Rama’s et al. (2012) investigations of a MUVE and MMOG respec-
tively lay important groundwork for further studies exploring TSCMC as integrated within a larger
computer-mediated environment. Their studies are also among the few so far that have answered
Sauro’s (2011) call for research into pedagogical tools that are specic to SCMC. Also worthy of
further exploration is Collentine’s (2013) nding that more informative yet linguistically simple input
optimally promoted output complexity, which if replicable after controlling for prociency, may inform
input-oriented theories of SLA.
Interaction (k = 3). Tare et al. (2014) investigated the eect of interaction on output and language
acquisition, comparing two classes: one which completed information gap tasks through TSCMC
versus one which worked on self-study versions of the tasks. Based on the Interaction Hypothesis
(Long, 1985; Gass & Mackey, 2007), interaction was postulated to facilitate more language acqui-
sition. Learners completed three approximately 20-minute assignments per week for six weeks as
homework. The interactive group produced signicantly more output, achieved signicantly higher
pre-post vocabulary gains, and demonstrated a signicant pre-post increase in production on an oral
test. TSCMC featured lexical scaolding in which one learner encountered target vocabulary used
by the other and engaged in NoM if it was not understood, possibly leading to the higher vocabulary
gains. Meanwhile, oral production may have increased due to the similarity of TSCMC to speaking
and the increased output on the assignments.
Golonka, Tare, and Bonilla (2017) further investigated chat data from Tare et al. (2014) to uncover
patterns of interaction that may have led to language gains, which included self- and peer-correc-
tion as well as NoM. NoM, though, was infrequent, whereas responding positively and self- and
peer-correction were more common. Golonka and colleagues postulate that the lack of NoM was due
to partners being of similar prociency. Regardless, this interaction produced signicant language
gains compared to individual study, suggesting that these other patterns may be as important for
language learning as NoM.
Yanguas (2020) compared the eects of L1 versus L2 TSCMC while working on a collaborative writ-
ing assignment in which learners narrated a picture sequence. The experimental study also included a
group that collaborated on the assignment but were not allowed to chat, as well as a control group that
did the assignment individually. A signicant dierence in accuracy emerged in favor of the L1 group,
which was able to take more than twice as many chat turns per minute as the L2 group and thereby
exchange more information. Their chat content also diered: whereas the L1 group focused more on
task planning, the L2 group devoted more attention to FoF. The L2 group also engaged in NoM, yet no
signicant eect on their writing emerged. These ndings may support Skehan’s Trade-O Hypothesis
(1998; 2009) with the L1 chat group experiencing lower cognitive load while writing because they did
not need to devote extra attention to simultaneously conversing in the target language. Meanwhile their
smooth L1 communications could provide more scaolding for ecient task completion, freeing up
further attention to devote to accuracy.
While Tare et al. (2014) presents evidence for the ecacy of TSCMC compared to individual study,
the follow-up analysis by Golanka et al. (2017) found that the most common features of interaction
related to language development were correction and the maintenance of positive aect rather than
NoM. Yanguas (2020) may further call into question the interactionist emphasis on NoM as well as the
cognitivist emphasis on FoF in the target language, with the nding that interaction in the L2 during
a collaborative writing assignment led to no signicant improvement in learners’ writing despite fea-
turing NoM and frequent FoF. Rather, learners who used their L1 during the writing task signicantly
outperformed the other groups.
11 Technology in Language Teaching & Learning, 4(1)
Grouping (k = 9)
Nine studies concerned group formation either by prociency and native/non-native speaker status or
by partner familiarity.
Prociency and status. Van der Zwaard and Bannink (2014) investigated interaction between native
(NS) and non-native speakers (NNS) and the eect of modality on NoM. Learners collaborated through
video-based SCMC or TSCMC to devise a dramatic scene incorporating jokes specic to the culture
of the NS. Given the NNSs’ unlikelihood of understanding the jokes initially, communication required
indicating non-understanding to initiate NoM for successful task completion (Varonis & Gass, 1985).
However, learners may avoid indicating non-understanding to save face (see Brown & Levinson, 1987
and Goman, 1967), especially when time limitations and nonverbal cues pressure them to respond
quickly and when the interlocutor is perceived as an expert on the topic. Indeed, NoM was more
stunted during video interaction where nonverbal cues and higher time pressure were present, than in
TSCMC which oered (citing Fox, 2004, p. 153) “the illusion of anonymity.”
Fredriksson’s (2015) investigated TSCMC among student triads in a German literature class involving
NSs and intermediate and advanced NNSs. NNSs tended to contribute less to the conversation when
their group included a NS. Furthermore, intermediate NNSs contributed more words and clauses when
the majority of group members were also intermediate NNSs as opposed to advanced. NoM occurred
rarely regardless of grouping with learners more commonly utilizing self-repair.
Eslami and Kung’s (2016) study on LREs during communicative tasks through TSCMC compli-
cates the notion that status inequality impedes communication. The researchers randomly paired
high- and low-level NNSs with a NS or with each other. Statistical analysis revealed that there was
no signicant dierence between the groups on the number of LREs initiated, but there was a signif-
icant dierence in the number of correctly resolved LREs in favor of the NNS-NS groups. Post-tests
showed no signicant dierence between groups in learning due to LREs. The researchers conclude
that pairing NNSs who have a large prociency gap between them can result in a similar amount
of LREs and learning as pairing NNSs with NSs. Regarding face-saving, the large prociency gap
between partners in the NNS-NNS pairs might have created a similar social dynamic of ‘novice’ and
‘expert’ between the NNSs that would have been operating among NNS-NS pairs. Note also that
the NSs were language teachers in training and thus likely engaged in LREs actively with NNSs in
order to ‘teach’ them.
Michel and O’Rourke (2019) focused on structural alignment and investigated learner-tutor versus
learner-learner interaction in TSCMC. The mean instances of alignment for learner-tutor interaction
were slightly higher and eye xations on n-grams that were aligned to occurred more often and lasted
longer for n-grams produced by tutors compared to peers. Alignment occurred rarely for both groups
however and no inferential statistics were performed.
Liu’s (2017) study involved NS teachers from an outside school and categorized learners as high (H)
and low (L) prociency, assigning them to H-NS, L-NS, H-H, or L-L pairs. Pairs interacted through
TSCMC to answer questions about short texts. Signicant dierences in the amount of NoM emerged
between groupings in this order (ranked highest to lowest): H-H, L-NS, H-NS, and L-L, with H-H
pairings resolving signicantly more NoMs than the others, whereas L-L pairings resolved signi-
cantly fewer. Thus, whether being paired with a NS led to more or less NoM depended on learners’
prociency. Note that the high prociency learners were Applied English majors likely studying to be
teachers, so they may have experienced face preservation issues when interacting with the NS teachers.
Hughes: Text-Based SCMC for SLA: A Narrative Review 12
Torres and Cung (2019) investigated interaction between 46 learners of Spanish as a heritage language
(HL) and 14 learners of Spanish as a second language (L2) assigning participants to L2-HL or HL-HL
pairings. Pairs completed two decision making and writing tasks through TSCMC and FTFC. FTFC
produced signicantly more LREs regardless of pairing and despite taking about half the time. Inter-
estingly, whereas the L2-HL pairs also resolved signicantly more LREs through FTFC, no signicant
dierence emerged between modes for HL-HL pairs. Thus, the childhood exposure of HL-HL pairs to
the target language and their consequent higher prociency may have allowed them to more eectively
adapt to TSCMC.
Finally, Coyle and Reverte (2017) conducted one of the few investigations involving younger partici-
pants who were aged 9 to 10. The tandem study assigned learners to lower-level (L-L) and higher-level
(H-H) pairings and had them do jigsaw story sequencing tasks in English through TSCMC (and tasks
in Spanish that were outside the scope of the study). H-H pairs used signicantly more negotiation
strategies than L-L pairs while neither group demonstrated substantial uptake. Whereas pairs were
technically composed of NNS-NS learners, the results align more with Lui’s (2017) H-H and L-L
pairings than that study’s H-NS and L-NS pairings. The reason may be that partners likely saw each
other as equals, as they did tasks in both of their respective target languages and those tasks were free
of highly culture-specic content. The participants were also children and may therefore have been
less concerned about face.
In summary, Van der Zwaard and Bannink (2014) and Fredriksson (2015) suggest that NS involve-
ment may hinder interaction and NoM due to issues of face. However, Eslami and Kung (2016)
and Michel and O’Rourke (2019) found that NNS-NS interaction produces more resolved LREs
and alignment respectively, while Lui (2017) discovered that NS involvement increased NoM for
low prociency NNSs but decreased it for high prociency NNSs. Meanwhile, Coyle and Reverte’s
(2017) NNS-NS pairings were more similar in their interaction patterns to Lui’s (2017) NNS-NNS
pairings due ostensibly to their perceiving each other as equals. In contrast, although the NSs in
Van der Zwaard and Bannink (2014) and Fredriksson (2015) were also fellow students, the tasks in
both studies involved highly culture-specic content which may have led other learners to see those
NSs as experts not only at the target language but also on the topic at hand leading to face-saving
strategies and decreased NoM. Likewise, Lui’s (2017) high prociency NNSs were likely English
teachers in training who when paired with NS teachers of English may have experienced face pres-
ervation issues due to a conict in self-perceived roles. Van der Zwaard and Bannink (2014) show
that, in such cases where face is likely to be an issue, TSCMC provides a sense of anonymity which
may ease face-saving and increase NoM. Finally, the ndings of Torres and Cung (2019) suggest
that high prociency learners may be able to adapt to the TSCMC mode more eectively than lower
prociency learners.
Partner familiarity. Dao, Duong et al. (2021) and Dao, Nguyen, Duong, and Tran-Thanh (2021)
investigated the eect of pairing learners with familiar versus unfamiliar partners. The study by Dao,
Duong et al. (2021), which has already been discussed in the Uptake, Noticing, and Alignment sec-
tion, sought to determine whether higher partner familiarity would lead to more frequent provision of
CF (referred to as ‘peer feedback’ by the researchers), but no signicant eect emerged. A similarly
designed study by Dao, Nguyen et al. (2021) involved 98 learners from private English schools in
Vietnam with a mean age of 16.93 and investigated learner engagement and its eect on the quality
of a subsequent collaborative writing assignment. Familiar partners demonstrated signicantly higher
learner engagement compared to unfamiliar partners, and regression analysis showed that measures of
learner engagement signicantly predicted the CAF of the written assignment. With regard to modal-
ity, all measures of learner engagement were signicantly higher for the video-based SCMC condition
compared to the TSCMC condition.
13 Technology in Language Teaching & Learning, 4(1)
In summary, whereas partner familiarity did not aect the provision of CF (Dao, Duong et al., 2021),
it did signicantly increase learner engagement (Dao, Nguyen et al., 2021), several measures of which
were then found to be predictors of the quality of a subsequent collaborative written assignment. These
results support the ecacy of allowing learners to partner with familiar peers for TSCMC tasks. The
fact that learner engagement was signicantly lower for TSCMC in the study compared to FTFC also
implies the need for measures to foster greater learner engagement during TSCMC tasks.
Conclusion
The analysis conducted here provides some tentative answers to the two questions posed by this review.
1. What aspects of SLA appear to be enhanced or hindered by TSCMC?
Beginning with hindrances, TSCMC is relatively time consuming with every study comparing TSCMC
to other modalities—except Kim’s (2017) and those that controlled for time—nding TSCMC to take
approximately twice as long or longer than spoken modalities while usually producing about half
the output. Also, at least in the case of learner-learner interaction CSs, NoM, CF, and uptake seem to
occur less in TSCMC compared to spoken modalities, although these dierences may disappear when
output amount is controlled for. Furthermore, despite its similarities to spoken language, learners may
need orientation and training in order to eectively utilize TSCMC for language learning tasks, and
TSCMC may be particularly problematic for non-alphabetic target languages, unless learners have
thorough prior experience typing in the language. Finally, it may be harder to foster learner engage-
ment in TSCMC than in spoken modalities, perhaps due to its lack of nonverbal ques and the time
delay between messages.
If enough time for orientation and task completion is provided, TSCMC may enhance SLA in several
ways. The permanent, textual nature of TSCMC as well as the time delay, may promote more noticing,
alignment, and the correct resolution of LREs compared to other modalities. It may also improve ques-
tion production and grammatical accuracy and lends itself more readily than spoken modes to delayed
CF. Additionally, TSCMC can provide the learner a sense of anonymity which may alleviate avoidance
of NoM due to face-preservation issues.
2. What are the insights provided and questions raised with regard to SLA theory and practice?
Studies on the eects of group formation suggest a complex dynamic between prociency and per-
ceived roles. Depending on the prociency, context, and the cultural specicity of target content, NoM
may be avoided in favor of face-saving strategies. TSCMC was also found to alleviate face preserva-
tion issues compared to FTFC. In addition, research on partner familiarity found it led to higher learner
engagement which in turn predicted learning, and maintenance of positive aect was one aspect of
learner engagement which Golanka et al. (2017) also found to be a frequent feature of interaction that
led to language gains in their study. These ndings may be particularly relevant to the sociocultural
account of SLA.
Informing the cognitive-interactionist perspective, studies investigating NoM found that it occurred
rarely during TSCMC, yet several observed signicant language gains. This suggests that other
features of interaction besides NoM played a larger role in learning. Tore et al. (2014) and Golanka
et al. (2017) found potential candidates in self- and other-correction as well as vocabulary model-
ling, while Smith (2012) showed that CF that was noticed led to language gains, providing support
for Schmidt’s Noticing Hypothesis (1990). Meanwhile, studies on structural alignment, which is
presumed to occur unconsciously, found that it too led to language development. However, these
other features of interaction also occur rarely in natural communication indicating the need for
further research.
Hughes: Text-Based SCMC for SLA: A Narrative Review 14
Studies on task design and conditions found that tasks featuring large amounts of input or extensive con-
text combined with minimal reasoning or syntactic processing requirements led to optimal language per-
formance in terms of output CAF and the resolution of linguistic problems. These ndings may support the
Trade-O Hypothesis and perhaps Krashen’s (1980) Input Hypothesis. However, Baralt’s (2013) ndings
complicate the picture, suggesting a possible interaction between cognitive complexity and modality, with
TSCMC operating according to the Trade-O Hypothesis while FTFC aligns more with the Cognition
Hypothesis. This interaction, if replicable, would have important implications for SLA theory and practice.
Finally, few studies have answered Sauro’s (2011) call for investigating pedagogical tools that are
specic to SCMC, with Peterson (2012) and Rama et al. (2012) conducting the only studies that inves-
tigated TSCMC as integrated within a larger computer-mediated environment and implying much
unexplored potential for this higher level of CMC. Future research might experiment not only with
adding mediation provided by a virtual environment, but with the direct mediation of learner-learner
interaction by the environment. Such dynamic CMC might enhance learner interaction by, for exam-
ple, making recommendations to interlocutors based on the content of their messages and would be of
interest to both the sociocultural and interactionist research paradigms.
This review inherits two main limitations from its narrative approach. First, in order to provide a su-
cient amount of detail on each study to enable in-depth comparison, the review draws from a slice of the
available literature on TSCMC rather than attempting exhaustive inclusion, and as with Sauro (2011), only
reviewed peer-reviewed journal articles, excluding other sources such as conference proceedings, book
chapters, and graduate theses. Second, while the emphasis on narrative over systematicity enabled the
elucidation of new insights into TSCMC for SLA, this approach may have been more suspectable to bias
and error on the part of the author. Also, this review only incorporated studies that were written in English.
Still, the review has showcased the potential of TSCMC for language learning and research. The results
have important practical and theoretical implications and provide various avenues for future exploration.
Disclosure Statement
The author reports there are no competing interests to declare.
References
Brown, P., & Levinson, S. C. (1987). Politeness. Some universals in language use [1978]. Cambridge
University Press.
Du, P. A., Norris, J. M., & Ortega, L. (2007). The future of research synthesis in applied linguistics:
Beyond art or science. TESOL Quarterly, 805–815. https://doi.org/10.1002/j.1545-7249.2007.
tb00105.x
Fox, K. (2004). Watching the English. London: Hodder and Stoughton Ltd.
Gass, S. M., & Mackey, A. (2007). Input, interaction, and output in second language acquisition. In
B. VanPatten & J. Williams (Eds.), Theories in second language acquisition: An introduction
(pp. 175–199). Mahwah, NJ: Lawrence Erlbaum.
Goman, E. (1967). Interaction ritual. Essays in face-to-face behavior. New Jersey: Transaction
Publishers.
Krashen, S. (1980). The input hypothesis. London: Longman.
Lin, H. (2014). Establishing an empirical link between computer-mediated communication (CMC)
and SLA: A meta-analysis of the research. Language Learning & Technology, 18(3), 120–147.
https://dx.doi.org/10125/44387
Lin, H. (2015). Computer-mediated communication (CMC) in L2 oral prociency development: A
meta-analysis. ReCALL, 27(3), 261–287. https://doi.org/10.1017/S095834401400041X
15 Technology in Language Teaching & Learning, 4(1)
Lin, W. C., Huang, H. T., & Liou, H. C. (2013). The eects of text-based SCMC on SLA: A meta
analysis. Language Learning & Technology, 17(2), 123–142. https://dx.doi.org/10125/44327
Long, M. (1985). Input and second language acquisition theory. In S. Gass & C. Madden (eds.). Input
in second language acquisition (pp. 377–393). Rowley, MA: Newbury House.
Ohta, A. (1995). Applying sociocultural theory to the analysis of learner discourse: learner learner
interaction in the zone of proximal development. Issues in Applied Linguistics, 6(2), 93–122.
https://doi.org/10.5070/L462005219
Robinson, P. (2001). Task complexity, cognitive resources and syllabus design: A triadic framework
for examining task inuences on SLA. In P. Robinson (Ed.), Cognition and second language
instruction (pp. 185–316). New York: Cambridge University Press.
Robinson, P. (2011). Second language task complexity, the Cognition Hypothesis, language learning, and
performance. In P. Robinson (Ed.), Second language task complexity: Researching the Cognition
Hypothesis of language learning and performance (pp. 3–38). Amsterdam: Benjamins.
Sauro, S. (2011). SCMC for SLA: A research synthesis. CALICO Journal, 28(2), 369–391. https://doi.
org/10.11139/cj.28.2.369-391
Schmidt, R. (1990). The role of consciousness in second language learning. Applied Linguistics, 11(2),
129–158. https://doi.org/10.1093/applin/11.2.129
Skehan, P. (1998). A cognitive approach to language learning. Oxford University Press.
Skehan, P. (2009). Modelling second language performance: Integrating complexity, accuracy, uency,
and lexis. Applied Linguistics, 30, 510–532. https://doi.org/10.1093/applin/amp047
Smith, B. (2017). Technology‐enhanced SLA research. In C. A. Chapelle & S. Sauro (Eds.), The
handbook of technology and second language teaching and learning (pp. 444–458). Oxford,
UK: Wiley Blackwell. https://doi.org/10.1002/9781118914069.ch29
Smith, B., & Laord, B. A. (2009). The evaluation of scholarly activity in computer-
assisted language learning. The Modern Language Journal, 93, 868–883. https://doi.
org/10.1111/j.1540-4781.2009.00978.x
Varonis, E., & Gass, S. (1985). Non-native/non-native conversations: A model for negotiation of mean-
ing. Applied Linguistics, 6, 71–90. https://doi.org/10.1093/applin/6.1.71
Ziegler, N. (2016). Synchronous computer-mediated communication and interaction: A meta-
analysis. Studies in Second Language Acquisition, 38(3), 553–586. https://doi.org/10.1017/
S027226311500025X
Appendix
List of studies included in this review.
Ajabshir, Z. F. (2019). The eect of synchronous and asynchronous computer-mediated communi-
cation (CMC) on EFL learners’ pragmatic competence. Computers in Human Behavior, 92,
169–177. https://doi.org/10.1016/j.chb.2018.11.015
Arroyo, D. C., & Yilmaz, Y. (2018). An open for replication study: The role of feedback timing in
synchronous computer-mediated communication. Language Learning, 68(4), 942–972. https://
doi.org/10.1111/lang.12300
Baralt, M. (2013). The impact of cognitive complexity on feedback ecacy during online versus face-
to-face interactive tasks. Studies in Second Language Acquisition, 35(4), 689–725. https://doi.
org/10.1017/S0272263113000429
Bower, J., & Kawaguchi, S. (2011). Negotiation of meaning and corrective feedback in Japanese/
English eTandem. Language Learning & Technology, 15(1), 41–71. http://dx.doi.
org/10125/44237
Brandl, K. (2012). Eects of required and optional exchange tasks in online language learning envi-
ronments. ReCALL, 24(1), 85–107. https://doi.org/10.1017/S0958344011000309
Hughes: Text-Based SCMC for SLA: A Narrative Review 16
Collentine, K. (2013). Using tracking technologies to study the eects of linguistic complexity in CALL
input and SCMC output. CALICO Journal, 30(1), 46–65. https://doi.org/10.1558/cj.v30i0.46-65
Coyle, Y., & Reverte, M. J. (2017). Children’s interaction and lexical acquisition in text-based online
chat. Language Learning & Technology, 21(2), 179–199. https://dx.doi.org/10125/44617
Dao, P., Duong, P. T., & Nguyen, M. X. N. C. (2021). Eects of SCMC mode and learner familiarity
on peer feedback in L2 interaction. Computer Assisted Language Learning, 1–29. https://doi.
org/10.1080/09588221.2021.1976212
Dao, P., Nguyen, M. X. N. C., Duong, P. T., & Tran-Thanh, V. U. (2021). Learners’ engagement in
L2 computer-mediated interaction: Chat mode, interlocutor familiarity, and text quality. The
Modern Language Journal. https://doi.org/10.1111/modl.12737
Darhower, M. A. (2014). Synchronous computer-mediated dynamic assessment: a case study
of L2 Spanish past narration. Calico Journal, 31(2), 221–243. https://doi.org/10.11139/
cj.31.2.221-243
Eslami, Z. R., & Kung, W. T. (2016). Focus-on-form and EFL learners’ language development in syn-
chronous computer-mediated communication: task-based interactions. The Language Learning
Journal, 44(4), 401–417.
Fredriksson, C. (2014). The inuence of group formation on learner participation, language complex-
ity, and corrective behaviour in synchronous written chat as part of academic German studies.
ReCALL, 27(2), 217–238. https://doi.org/10.1017/S0958344014000238
Gigure, C., & Parks, S. (2018). Child-to-child interaction during eTandem ESL-FSL chat exchanges.
Language Learning & Technology, 22(3), 176–192. https://dx.doi.org/10125/44663
Golonka, E. M., Tare, M., & Bonilla, C. (2017). Peer interaction in text chat: Qualitative analysis of chat
transcripts. Language Learning & Technology, 21(2), 157–178. https://dx.doi.org/10125/44616
Hsu, H. C. (2012). Investigating the eects of planning on L2 text-chat performance. CALICO Journal,
29(4), 619–638. https://doi.org/10.11139/cj.29.4.619-638
Hsu, H. C. (2015). The eect of task planning on L2 performance and L2 development in text-based
synchronous computer-mediated communication. Applied Linguistics, 38(3), 359–385. https://
doi.org/10.1093/applin/amv032
Hung, Y. W., & Higgins, S. (2015). Learners’ use of communication strategies in text-based and
video-based synchronous computer-mediated communication environments: Opportunities for
language learning. Computer Assisted Language Learning, 29(5), 901–924. https://doi.org/10.
1080/09588221.2015.1074589
Kessler, M., Polio, C., Xu, C., & Hao, X. (2020). The eects of oral discussion and text chat on L2
Chinese writing. Foreign Language Annals, 53(4), 666–685. https://doi.org/10.1111/an.12491
Kim, H. Y. (2012). Learning opportunities in synchronous computer-mediated communication and
face-to-face interaction. Computer Assisted Language Learning, 27(1), 26–43. https://doi.org/
10.1080/09588221.2012.692386
Kim, H. Y. (2017). Eect of modality and task type on interlanguage variation. ReCALL, 29(2),
219–236. https://doi.org/10.1017/S0958344017000015
Kim, Y., Jung, Y., & Skalicky, S. (2019). Linguistic alignment, learner characteristics, and the produc-
tion of stranded prepositions in relative clauses: Comparing FTF and SCMC contexts. Studies
in Second Language Acquisition, 41(5), 937–969. https://doi.org/10.1017/S0272263119000093
Kim, Y., Skalicky, S., & Jung, Y. (2020). The role of linguistic alignment on question development
in face-to-face and synchronous computer-mediated communication contexts: A conceptual
replication study. Language Learning, 70(3), 643–684. https://doi.org/10.1111/lang.12393
Kourtali, N. E. (2022). The eects of face-to-face and computer-mediated recasts on L2 development.
Language Learning & Technology, 26(1), 1–20. https://dx.doi.org/10125/73457
Liu, S. H. J. (2017). Text-based negotiated interaction of NNS-NNS and NNS-NS dyads on Facebook.
ReCALL, 29(3), 294–312. https://doi.org/10.1017/S0958344017000143
17 Technology in Language Teaching & Learning, 4(1)
Michel, M., & O’Rourke, B. (2019). What drives alignment during text chat with a peer vs. a tutor?
Insights from cued interviews and eye-tracking. System, 83, 50–63. https://doi.org/10.1016/j.
system.2019.02.009
Peterson, M. (2012). EFL learner collaborative interaction in Second Life. ReCALL, 24(1), 20–39.
https://doi.org/10.1017/S0958344011000279
Rama, P. S., Black, R. W., Van Es, E., & Warschauer, M. (2012). Aordances for second lan-
guage learning in World of Warcraft. ReCALL, 24(3), 322–338. https://doi.org/10.1017/
S0958344012000171
Rouhshad, A., Wigglesworth, G., & Storch, N. (2016). The nature of negotiations in face-to-face versus
computer-mediated communication in pair interactions. Language Teaching Research, 20(4),
514–534. https://doi.org/10.1177/1362168815584455
Smith, B. (2012). Eye tracking as a measure of noticing: A study of explicit recasts in SCMC. Language
Learning & Technology, 16(3), 53–81. https://dx.doi.org/10125/44300
Tare, M., Golonka, E. M., Vatz, K., Bonilla, C. L., Crooks, C. & Strong, R. (2014). Eects of interac-
tive chat versus independent writing on L2 learning. Language Learning & Technology, 18(3),
208–227. https://dx.doi.org/10125/44391
Torres, J., & Cung, B. (2019). A comparison of advanced heritage language learners’ peer interaction
across modes and pair types. The Modern Language Journal, 103(4), 815–830. https://doi.
org/10.1111/modl.12594
Uzum, B. (2010). An investigation of alignment in CMC from a sociocognitive perspective. CALICO
Journal, 28(1), 135–155. https://doi.org/10.11139/cj.28.1.135-155
Van der Zwaard, R., & Bannink, A. (2014). Video call or chat? Negotiation of meaning and issues of
face in telecollaboration. System, 44, 137–148. https://doi.org/10.1016/j.system.2014.03.007
Yanguas, I. (2019). L1 vs L2 synchronous text-based interaction in computer-mediated L2 writing.
System, 88, 102169. https://doi.org/10.1016/j.system.2019.102169
Yilmaz, Y. (2011). Task eects on focus on form in synchronous computer‐ mediated
communication. The Modern Language Journal, 95(1), 115–132. https://doi.org/10.1111/
j.1540-4781.2010.01143.x
Yuksel, D., & Inan, B. (2014). The eects of communication mode on negotiation of meaning and its
noticing. ReCALL, 26(3), 333–354. https://doi.org/10.1017/S0958344014000147
Zeng, G. (2017). Collaborative dialogue in synchronous computer-mediated communication and face-to-
face communication. ReCALL, 29(3), 257–275. https://doi.org/10.1017/S0958344017000118
Ziegler, N. (2018). Pre-task planning in L2 text-chat: Examining learners’ process and performance.
Language Learning & Technology, 22(3), 193–213. https://dx.doi.org/10125/44664
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
This study investigated the effects of synchronous computer-mediated communication (SCMC) mode and interlocutor familiarity on frequency and characteristics of peer feedback in L2 interaction. Fifty dyads of EFL learners were equally assigned into familiar (+/–) groups and performed an interactive task in two SCMC modes (text/video-chats). After their interactions, they were interviewed individually about the impact of SCMC mode and interlocutor familiarity on the provision of feedback. Learners’ text/video-chats were coded for feedback frequency and characteristics (e.g. type, linguistic focus, accuracy, and modified output). Results show that more instances of feedback were observed in the video- than text-chats; however, interlocutor familiarity did not affect the amount of feedback. Despite differences in types, feedback’s linguistic focus and accuracy, frequency and characteristics of modified output were relatively similar between two SCMC modes. Content-based analyses of the interviews revealed that learners attributed the differences in feedback occurrence to various characteristics of the SCMC modes rather than interlocutor unfamiliarity. The results suggest greater benefits of the video-chat over the text-chat in promoting peer feedback and emphasise the importance of establishing a positive relationship among learners during L2 SCMC interaction.
Article
Full-text available
This study investigated the impact of synchronous computer‐mediated communication (SCMC) mode and familiarity with partners on learner engagement in second‐language task‐based interaction, and whether learner engagement is linked to subsequent joint‐written‐text quality. Ninety‐eight Vietnamese learners of English were assigned into (±) familiar groups and performed a picture‐sequencing tasks in 2 SCMC modes (i.e., video and text chat). Scores of 3 types of learner engagement (cognitive, social, and emotional) were compared across the conditions. Results showed that scores of all engagement types in the video chat were significantly higher than in the text chat. Familiar dyads also showed higher engagement than unfamiliar peers during the interaction. Learners reported different reasons for their preferences of video chat over text chat. Language‐related episodes, semantically engaged talk, and mutual help as measures of learner engagement were predictive of the subsequent text quality. The results contribute to the general understanding of the characteristics of video and text chat and their impact on learner engagement and text quality.
Article
Full-text available
The role of recasts, a corrective feedback technique, has received much attention from instructed SLA researchers. While a variety of factors have been identified as influencing their effectiveness in facilitating uptake and L2 development (e.g., learners' age and level of proficiency), the role of mode of interaction has been the object of relatively little research. To fill this gap, the current study explored the impact of mode of interaction on learners' successful uptake and L2 gains when recasts are provided. Sixty young Greek EFL learners (M = 11.39 years old, SD = .86) were assigned to one of two experimental conditions that differed as to whether students engaged in synchronous computer-mediated communication (SCMC) or face-to-face (FTF) interaction. Both groups performed information transmission tasks that required them to provide information about habits of fictional characters. In both conditions, the participants received interrogative, partial recasts addressing errors on the present third person singular. The recasts were oral in the FTF condition and written in the SCMC condition. L2 development was gauged by an oral and a written production test. Results demonstrated that oral recasts in the FTF mode generated more successful uptake and they led to more L2 gains than written recasts in the SCMC condition on both outcome measures.
Article
Full-text available
Numerous studies have investigated the effects of prewriting tasks on writing quality and language (e.g., individual vs. collaborative planning, text-chat vs. face-to-face [FTF] discussion, etc.). Recent meta-analyses have found a small advantage for text chat on writing, but no studies involved logographic languages, for which character learning may limit and tax language production. In this study we explore this issue, as university learners of Chinese (N = 10) engaged in prewriting discussions orally and via text chat followed by a timed-writing task. Results show that during FTF planning, students engaged in talk that was greater in terms of discussion length and number of turns; in students' writing, FTF planning resulted in increased lexical complexity and syntactic richness. Our results contradict theoretical perspectives suggesting that text chat might be more advantageous, as well as one empirical study by Liao (2018). Finally, we discuss the implications for learning to write in a non-alphabetic language.
Article
To date, linguistic alignment studies in second language acquisition have mainly been conducted during face‐to‐face (FTF) interactions. In the current study, we examined and compared the effect of structural alignment on the development of English direct and indirect questions in FTF and synchronous computer‐mediated communication (SCMC) contexts. Additionally, we assessed participants’ working memory capacity and previous knowledge of the target structures. We assigned 50 Korean learners of English to either an FTF or SCMC modality. Over 3 weeks, they completed a pretest, two interactive alignment sessions, a working memory test, and immediate and delayed posttests. Results indicate that primed production of direct and indirect questions is significantly more likely than unprimed production. Moreover, we found an interaction with modality for primed production of direct questions but not for indirect questions. Findings suggest learning effects for direct questions only and that this effect is facilitated by the degree of alignment.
Article
A strand of task‐based interaction research has emerged to better understand the effects of heritage language (HL) and second language (L2) learners’ peer collaboration on interactional moves (e.g., language‐related episodes [LREs], self‐repairs) and linguistic focus. To extend this line of research, this study compared 14 HL–L2 and 16 HL–HL advanced learners’ peer interactions across face‐to‐face (FTF) and written synchronous computer‐mediated communication (SCMC) modes. Each dyad completed 2 decision‐making/collaborative writing tasks in Spanish across both interaction modes. Results revealed that interaction mode had a large effect on the extent to which HL–L2 and HL–HL pairs produced self‐repairs and initiated LREs. That is, both pair types initiated more LREs in FTF mode, and self‐repaired non‐target‐like utterances more extensively in SCMC mode. As for pair types, HL–L2 pairs significantly resolved more lexis‐focused episodes in FTF mode, and HL–HL pairs only produced self‐repair episodes in SCMC mode. HL–L2 and HL–HL dyads addressed linguistic items (e.g., morphosyntax) rather equally across interaction modes. Our findings are discussed in light of the differences in prior language‐learning experiences of HL and L2 learners.
Article
This study investigates the quality of the L2 written outcome and the interactions produced in synchronous text-based L1 and L2 collaboration. On the one hand, it explores fluency, accuracy, and complexity measures in order to compare participants’ L2 performance in a task-based writing assignment using Google Docs. On the other hand, it examines the areas of negotiation in their L1 and L2 synchronous interactions in order to assess possible differences. Eighty-five intermediate students of Spanish were randomly assigned to four different dyadic writing groups: a control group, a collaborative L1 text-based chat group, a collaborative L2 text-based chat group, and a group of dyads who collaborated on Google Docs but were not permitted to chat. ANOVA analyses of the L2 writing measures under scrutiny showed that the dyads that collaborated in their L1 had a significant advantage in the accuracy measure utilized. Additionally, analyses of the interactions produced by the dyads in both chat groups showed that the L1 chat group conversed more but focused less on the L2. These results are discussed in relation to the relevant previous literature and should continue the discussion on translanguaging and how L1 can be of use in L2 writing.
Article
The current study examined the occurrence and benefits of linguistic alignment in two modalities, face-to-face (FTF) and synchronous computer-mediated communication (SCMC), focusing on stranded prepositions in relative clauses. It further examined how learner characteristics (i.e., working memory, language proficiency, previous knowledge of the target structure) mediate the effects of linguistic alignment. Ninety-four Korean students were assigned to one of the following groups: FTF alignment, SCMC alignment, FTF control, and SCMC control. The alignment experimental groups completed two alignment sessions, finished three stranded preposition tests, and carried out a running span test and cloze test over three weeks. Results indicated not only that linguistic alignment occurred in both FTF and SCMC modes but also that alignment was facilitated significantly more in the SCMC than FTF interactions. Furthermore, the findings suggest immediate and delayed learning effects in both modalities, and that learners’ prior knowledge of the target structure was significantly associated with the occurrence of alignment.
Article
Alignment refers to the largely automatic tendency of interlocutors to re-use each other’s language patterns in the course of authentic interaction (Pickering & Garrod, 2004). However, little is known about alignment during L2 interaction, where recycling of a partner’s language - in contrast to L1 conversations - might be a more conscious, strategic behaviour (Costa, Pickering & Sorace, 2008). In this paper we explore attention and awareness of L2 speakers to their partner’s language use during digitally mediated communication (text chat) and see whether and how they re-use each other’s lexical phrases. Twelve English-L1 learners of L2 German participated in a task-based text chat discussion. Participants communicated under two different conditions: with a native speaker tutor vs. with a language learning peer. Three data sources were scrutinised: (1) chatlog analyses looked for re-used lexical phrases; (2) eye-gaze fixations on re-used text served as measures of attention; (3) awareness of language re-use was evaluated by means of cued interviews. Results reveal both unaware and strategic alignment; more of it happening in the tutor chat; and more advanced students being less inclined to copy their partner. Findings will be discussed in light of Costa et al.’s (2008) framework on L2 alignment.
Article
Research suggests that pre-task planning time provides learners with opportunities to formulate, organize, and mentally store content, thereby freeing up attentional resources during tasks (Skehan, Xiaoyue, Qian, & Wang, 2012). However, relatively few studies to date have investigated pre-task planning in a synchronous computer-mediated communication setting (e.g., Lai, Fei, & Roots, 2008; Hsu, 2012, 2015). In addition to a scarcity of computer-assisted language learning research, relatively little is known about what learners do when they plan or how they use their plans during tasks. The goals of the current study were twofold: (a) to examine the relationship between pre-task planning and learners' production and (b) to explore the affordances offered by computer-mediated contexts to further investigate how and what learners may (or may not) be planning during pre-task and within-task planning time. Results suggest that three minutes of planning time resulted in increases in lexical complexity (but not phrasal or syntactic), although no significant findings were identified for accuracy or fluency. In addition, findings indicate that technology offers researchers a number of unique methodological affordances, such as the ability to see what learners produce, regardless of whether they transmit this information to their interlocutor, thereby providing evidence of L2 knowledge that would otherwise be unobservable.