SLA: Activities for Meaningful Interaction

1
SLA Activities for Meaningful Interaction
LING 561/761 09/13/2006
2
Computer-Mediated Language Learning Contexts (1)

• Research focusing directly on computer-assisted
  language learning has remained largely on the
  margins of SLA research (Chapelle, 2004
  Hulstijn, 2000) as it has struggled to
  contextualize itself within a larger theoretical
  framework of SLA (Chapelle, 1998, 2001 Doughty
  Long 2003 Salaberry, 2000).

3
Computer-Mediated Language Learning Contexts (2)

• Computer-Mediated Communication (CMC)
• Computer-Assisted Language Learning (CALL)
• Intelligent Computer-Assisted Language Learning
• (ICALL)
• Interaction occurs in a written modality.
• Discourse proceeds at a slower pace than in oral
  
• interaction.
• More time for preparing/repairing utterances.
• Turn-taking sequences/allocation of turns are
  unique and
• vary according to the specific
  medium.

4
What are the challenges? (1)

• For SLA Research
• Comparability to traditional SLA research
  paradigms.
• Can computer-centered language learning be
  task-based?
• Can attested results from oral interaction be
  brought to bear on human-computer interaction?
• Modality
• How do we account for the role of modality in
  this type of SLA research?
• Visual, text-based, written
• No prosodic or paralinguistic cues
• More time for online processing/planning
• Few to no socio-cultural constraints

5
What are the challenges? (2)

• For Language Pedagogy
• Traditional classrooms
• How are computer-centered tasks actually used in
  traditional language classes?
• How might they more effectively complement
  traditional teaching?
• What are the technical limitations? Can they be
  overcome?
• Distance education
• Same as above.

6
SLA Claims in CALL Research (1)

• "Because oral interaction is considered by many
  to be important for second language development,
  and because synchronous CMC bears a striking
  resemblance to oral communication, it seems
  logical to assume that language practice through
  CMC will reap some of the same benefits for
  second language development as practice through
  oral interaction."
• Pelletieri (2000, p. 59)

7
SLA Claims in CALL Research (2)

• The findings suggest that computer mediated
  communication (CMC) can provide many of the
  alleged benefits ascribed to the Interaction
  Hypothesis, but with greatly increased
  possibilities for access outside of the classroom
  environment.
• Blake (2000, p. 120)
• "Networked exchanges, since they are text-based
  and learners must type out or produce the
  structures in question, appear to constitute an
  example of forced output ."
• Blake (2000, p. 132)

8
Feedback in CMC Research (1)

• Claims from CMC research include
• Computer-mediated peer feedback is neither
  superior nor inferior to orally provided peer
  editing feedback in promoting second language
  writing development (Schultz, 2000).
• Learners negotiate for meaning, provide feedback
  and modify their output when engaged in CMC
  tasks, especially in response to lexical and
  structural difficulties (Tudini, 2003).

9
Feedback in CMC Research (2)

• Claims from CMC research (cont.)
• "Because CMC fosters negotiation of meaning and
  form-focused interaction and because students
  communicating through this medium have more time
  to process and monitor the interlanguage, I
  believe that CMC can play a significant role in
  the development of grammatical competence."
  (Pelletieri, 2000, p. 83)

10
Feedback in ICALL Research (1)

• Learners engage with computer programs that
  incorporate natural language processing (NLP)
  tools, such as syntactic parsers.
• NLP tools allow for online processing of
  responses and the provision of feedback that is
  tailored to the students needs and highly
  informative about the nature of their errors
  (Nagata Swisher, 1995).
• Tasks are highly artificial and bear little
  resemblance to naturally occurring oral
  interaction.

11
Feedback in ICALL Research (2)

• Claims from ICALL research include
• Metalinguistic feedback in ICALL systems promotes
  L2 grammatical development which is superior to
  (1) traditional (pre-scripted) computer feedback
  (Nagata, 1993 Nagata Swisher, 1995) and (2)
  translation feedback (Nagata, 1997).
• Metalinguistic feedback that highlights the error
  in the student input (i.e. input enhancement) is
  more effective at eliciting learner uptake than
  recasts input enhancement (Heift, 2004).

12
Recasts

• Recasts convey needed information about the
  target language in context, when the
  interlocutors share a joint attentional focus,
  and when the learner already has prior
  comprehension of at least part of the message,
  thereby facilitating form-function mapping.
  (Long, in press, p.46)
• The efficacy of recasts for promoting language
  development also lies in the immediate
  juxtaposition of the learners error and the
  correct reformulation provided by the more
  advanced speaker (Farrar, 1990 Long, 1996
  Saxton, 1997, 2005).

13
Recasts (2)

• It is claimed that this juxtaposition of learner
  error and subsequent reformulation
• 1. Enhances the salience of the corrected
  linguistic form in the feedback (Ishida, 2004
  Saxton, 1997)
• 2. Provides an opportunity for learners to make a
  cognitive comparison between the targetlike model
  in the recast and their own nontargetlike
  production, thus promoting restructuring of their
  interlanguage representation of the form (Long
  Robinson, 1998 Mackey Philp, 1998 Oliver,
  1995 Philp, 2003).

14
Recasts (3)

• These claims about the efficacy of recasts assume
  that the learner recognizes that the recasts are
  intended as corrective feedback.
• Although, it has been argued that
• Recasts are often not perceived as corrective
• feedback in some contexts (Lyster, 1998a
  Lyster,
• 1998b Lyster, 2004 Lyster Ranta, 1997).
• The implicit negative feedback contained in
  recasts
• may not be as crucial to L2 development as the
  
• enhanced salience of the positive evidence
  they
• provide (Leeman, 2003).

15
Recasts Salience (1)

• According to Schmidts (Schmidt, 1990, 1995,
  2001 Schmidt Frota, 1986) noticing
  hypothesis, in order to become potential
  candidates for intake and subsequent learning,
  formal linguistic features in the input must
  first be noticed by the learner (Schmidt, 1995,
  p. 20).
• When a learner's incorrect utterance and
  subsequent recast convey the same meaning and
  differ only in the use of a particular linguistic
  form, that form is said to be made perceptually
  more salient to the learner (Farrar, 1990 Long,
  in press)

16
Recasts Salience (2)

• Yet, the extent to which the increased salience
  of a linguistic form may be a function of a
  recast is constrained by a broad range of
  factors, such as
• The learners' developmental readiness vis-à-vis
  the corrected form (Han, 2002 Iwashita, 2003
  Mackey Philp, 1998 Philp, 2003)
• The length of the recast provided (Philp, 2003
  Sheen, 2004)
• The number of corrections made in the
  reformulation (Bohannon Stanowicz, 1988
  Farrar, 1992 Philp, 2003)
• The length of the negotiation sequence in which
  it occurs (Ellis, Basturkmen and Loewen, 2001)
• The degree of elaboration with which it is
  delivered (Doughty Varela, 1998)

17
Recasts Pedagogical Context (1)

• The salience and ambiguity of recasts are, to
  some extent, dependent upon the pedagogical
  context in which they appear (Nabei Swain,
  2002 Nicholas, Lightbown Spada, 2001 Oliver,
  1995 Sheen, 2004).
• In communicative, immersion language classrooms,
  where the instructional focus is primarily on
  meaningful language production, recasts and
  repetitions are often difficult for learners to
  disambiguate (Lyster, 1998a Lyster, 1998b,
  Lyster, 2004 Lyster Ranta, 1997)

18
Recasts Pedagogical Context (2)

• In more form-focused classroom environments (e.g.
  Ellis, Basturkmen Loewen, 2001 Ohta, 2000
  Sheen, 2004) and especially in dyadic laboratory
  contexts (e.g. Han, 2002 Iwashita, 2003 Long,
  Inagaki, and Ortega, 1998 Mackey Philp, 1998
  Philp, 2003) recasts may be more likely to be
  noticed as corrective feedback.
• One largely unexplored pedagogical context for
  recasts is in the emergent environment of
  interactive, computer-centered written discourse.

19
Recasts Pedagogical Context (3)

• As Long (2006) notes, there is a need for
  further research in this area because
• (1) the written modality is a robust
  environment for manipulating the degrees of
  saliency of target items and
• (2) findings from this line of research may have
  important pedagogical implications for teachers
  and materials developers in distance language
  instruction programs.

20
Recasts Pedagogical Context (4)

• Ayoun (2001) is the only published study to date
  to explore the effects of recasts on L2
  development in the written modality.
• Ayoun found that written recasts produced
  statistically significantly greater gain scores,
  compared to explicit grammatical instruction, in
  the aspectual distinction of distinction the
  French passe compose and imparfait.
• Relative to written models, exposure to written
  recasts produced greater, though not significant
  gains

21
Recasts Pedagogical Context (5)

• Ayoun's findings, however, are difficult to
  compare to findings on oral recasts for several
  reasons
• 1. Output was highly constrained and artificial.
• 2. Recasts were provided regardless of the
  presence of
• errors in the learners utterances.
• 3. No empirical claims were made about how the
  written
• modality may have impacted the
  effectiveness of recasts.

22
Overview

• This study will explore
•  
• The effects of interactional mode (written vs.
  oral) on the efficacy of recasts in promoting
  both short and long-term developmental gains in
  ESL question formation.
• The extent to which interactional context might
  influence learners' perceptions about the recasts
  provided in response to their errors.

23
Methodology Overview

• This study will employ a pre-test/post-test/delay
  ed post-test design with a series of experimental
  treatment tasks designed to isolate the context
  of interactional mode (oral interaction vs.
  computer-generated interaction) as the primary
  independent variable.

24
Participants

• Participants for this study will be recruited
  from the English as a Foreign Language Intensive
  Program at Georgetown University. All
  participants will be offered monetary
  compensation, and will be paid for each session
  that they attend.

25
Experimental Groups

• 1. Oral Interaction Recast Participants in this
  group will participate in dyadic, communicative
  tasks with a native speaker and receive intensive
  recasts in response to their errors related to
  question formation.
• 2. Computer Guided Interaction Recast
  Participants in this group will participate in
  communicative tasks on a computer and receive
  intensive recasts generated by the software in
  response to their errors related to question
  formation.
• 3. Control Participants in this group will only
  take the pre and post-tests.

26
Target Form

• ESL question formation
• Well attested in SLA research (e.g. Adams, 2004
  MacDonough, 2005 Mackey, 1999 Mackey
  Philp,1999 Pienemann, 1998 Pienemann
  Johnston, 1987 Pienemann, Johnston
  Brindley,1988 Philp, 2003 Silver, 2000 Spada
  Lightbown, 1993,1999 White, Spada, Lightbown
  Ranta, 1991)
• Well suited to focused feedback generated by NLP
  tools.

27
Developmental Sequences
28
Operationalizing Recasts (1)

• Only errors related to question formation will be
  corrected (targeting Stage 3-gt4 and Stage 4-gt5
  development).
• Lexical, phonological/typographic, false starts,
  hesitations and all other errors in the learners'
  speech and writing will not be subject to
  corrective feedback.
• No additional emphasis will be placed on the
  source of the error in either written or oral
  modes.
• Recasts will reformulate ill-formed utterances in
  their entirety (no partial or segmented recasts
  will be provided).

29
Operationalizing Recasts (2)

• Following Long et al. (1998) and Leeman (2003),
  no opportunity for modified output will be
  provided in either mode. In both the oral and
  computer guided interaction conditions, recasts
  will be followed by a prompt to continue.
• The decision to not allow for any uptake or
  repair is motivated primarily by the desire to
  control and isolate the effect of recasts from
  those of modified output.
• L2 development and learners' perceptions of
  recasts will be measured by gains from pre-test
  to post-test performance and reported noticing in
  stimulated recall protocols.

30
Experimental Schedule
31
Treatment/Assessment Tasks (1)

• Tasks will be counter-balanced and will include
  numerous contexts for Stage 4 and Stage 5
  questions to occur.
• Assessment tasks will be administered in both
  oral and written modes to all groups.
• Following the recommendation of Schneiderman
  (1992), the computer-guided tasks will be created
  in an anthropomorphic context to engage the
  learner in oral-like interaction.
• A virtual interlocutor will introduce each task,
  allocate turns and provide recasts in response to
  errors.

32
Treatment/Assessment Tasks (2)

• The oral and written tasks will include
• Spot the difference
• Picture sequencing
• Picture Matching

33
Example Task