Does Output Promote Noticing and Second Language Acquisition

1
Does Output Promote Noticing and Second Language
Acquisition?

• Izumi, S. Bigelow, M. (2000)

Presented by Eun Sung Park Ji Hyun Kim
2
The Output Hypothesis (Swain, 1995)

• Three functions of output
• 1) the hypothesis-testing function
• 2) the metalinguistic function
• (conscious reflection)
• 3) the noticing/triggering function
• The current study is concerned with

3
The Noticing Function of Output

• Producing the target language may prompt L2
  learners to consciously recognize some of their
  linguistic problems and make them aware of
  something they need to find out about their L2.

4
Noticing and SLA

• Schmidt (1990, 1994)
• Noticing at the level of awareness is a necessary
  and sufficient condition for converting input
  into intake.
• What must be attended to and noticed is not just
  the input in a global sense but whatever features
  of the input that are relevant for the target
  system.

5
Detection

• Tomlin Villa (1994)
• Detection is the process by which particular
  exemplars are registered in memory. Detection
  does not necessarily imply awareness.
• Awareness may be dissociated from attention and
  it is not necessary for learning.

6
Noticing vs. Detection

• Underlying the conflict between Schmidts
  noticing and Tomlin Villas detection
• is the role of __ __ __ __ __ __ __ __ __.
• (awareness)
• peacemaker
• Robinson (1995) Noticing is detection plus
  rehearsal in the short-term memory.

7
Is noticing measurable?

• Is noticing an observable behavior?
• How would you operationalize noticing?

8
Problems involved in measuring noticing

• Apart from the debate surrounding the role of
  awareness in learning, the measurement of
  noticing presents difficulties because it
  involves examining learner- i_ _ _ _ _ _ _ _
  processes.

9
How did previous investigations tackle the issue
of noticing/detection?

• Investigated the effects of some kind of external
  manipulation of input on learners
  noticing/detection and intake by employing post
  exposure measures
• Employed some type of online measures to address
  the issue of noticing/detection

10
Post-exposure Measures

• Grammaticality judgment tasks completed after
  exposure to different kinds of instruction such
  as input enhancement and input flood.
• (Alanen, 1995 Doughty, 1991 Trahey White,
  1993)
• Debriefing questionnaires or interviews conducted
  after the treatment period.
• (Alanen, 1995 Izumi et al., 1999)
• Stimulated recall while watching a video replay
  of an interaction. (Mackey et al., 1999)

11
Online measures

• Checking/marking different features of morphemes
  while listening to oral input.
• (VanPatten, 1990)
• Producing think-aloud protocols during a reading
  task or while solving linguistic puzzles.
• (Alanen, 1995 Leow, 1998)
• Metatalking during dictogloss.
• (Swain Lapkin, 1995)
• Underlining a grammatical feature embedded in
  written texts.
• (Izumi et al., 1999)

12
Post-exposure vs. online measures

• O_________ measures have an advantage over
  p____________ measures of noticing and detection
  because they allow more d_____________ access to
  learners ongoing i_____________ processes and
  m___________ possible memory loss.

13
In light of the growing evidence that attention
plays a central role in SLA, various pedagogies
have been proposed in order to draw the learners
attention to grammatical features in the input

• Consciousness-raising/input enhancement
• Focus on form
• These pedagogical proposals have been tested
  in various empirical studies grammar-based
  tasks, feedback tasks, or a combination of both,
  textual input enhancement tasks, recast studies,
  task-based interaction, input flood, processing
  instruction studies.

14
Collectively,

• The results of such studies suggest that drawing
  learners attention to form facilitates L2
  learning.

15
Output Noticing Hypothesis
vote
Input
Noticing the gap between their IL TL
Noticing problems
facilitate
This noticing may in turn the
processes of language acquisition. Moreover, if
output triggers attention to form, learners may
pay attention to and
simultaneously when they initiate production
with the intention of conveying content.  
meaning
form
16
Swain Lapkin (1995)
Research Question
Whether output would lead to conscious
recognition of problems and whether this
recognition would activate the kinds of cognitive
processes that are believed to be conducive to
SLA.
Findings
The learners do recognize problems in their IL
during production. Recognition of problems often
triggers cognitive processes that have been
implicated in L2 learning processes that either
generate linguistic knowledge that is new for
the learners or consolidate their existing
knowledge.
17
Swain Lapkin (1995)
Output not only leads to noticing of the gaps in
IL knowledge but also facilitates language
acquisition by triggering various internal
processes conducive to SLA.
However,
this study did not address the question of
whether the awareness of problems during
production can prompt the learner to seek out
subsequent input with more focused attention.
 
18
Izumi et al. (1999)
Research Question
Whether awareness of problems during production
would prompt L2 learners to seek out subsequent
input with more focused attention and lead to
the noticing and learning of a specific
grammatical form.
Experimental group (EG) Output Input
Text Reconstruction Task
Essay Writing Task
Comparison group (CG) Input
Comprehension Task
19
Izumi et al. (1999)
Findings
1. Both groups showed increased noticing of the
target form.
2. The experimental group demonstrated a
significant rate of uptake of the target form in
their production immediately following
exposure to the input.
3. The experimental group made significantly
larger gains on the production test after Phase 2
of the treatment than did the comparison group.
These results lend some support to the noticing
function of output in SLA.
20
Izumi et al. (1999)
Relative contribution of the two types of output
treatments
Reconstruction Task
Essay Writing
Lower median scores larger individual
variation
The impact of the first treatment or other
variables such as the nature of task
21
Izumi et al. (1999)
Relative contribution of the two types of output
treatments
Theoretically, this issue is relevant to the
noticing function of output. (i.e., do the
conditions under which output occurs somehow
constrain or mediate the effect of output on
noticing and learning or, more broadly, on the
learners attentional mechanism?)
Pedagogically, it is relevant to the question
concerning what kinds of tasks better facilitate
noticing and learning.
22
Izumi Bigelow (2000)
To explore the role of task ordering in the
effect of output on noticing and learning, the
researchers made one modification.
They reversed the order of the treatments.
Reconstruction Task
Essay Writing Task
Research Questions
1. Do output activates promote the noticing of
linguistic form in subsequent input?
2. Do these output-input activates result in
improved production of the target form?
23
Izumi Bigelow (2000)
Hypothesis 1

• The EG would show greater noticing of the target
  form contained in the input.

Hypothesis 2

• After being exposed to input containing the
  target form, the EG would indicate immediate
  uptake of the target form in their output.

Hypothesis 3

• The EG would show greater gains in accuracy of
  the use of the target form in the posttests.
•  

24
Method

• Participants
• 18 adult ESL learners enrolled in an academic
  writing class in a community college.
• Experimental Group (EG) 9 students
• Comparison Group (CG) 9 students

25
Form in Focus

• The past hypothetical conditional
• e.g., If Ann had traveled to Korea last summer,
  she would have seen the World Cup.
• This structure was not completely new to the
  students, but most of them were not able to use
  it accurately.
• Examples of students production of the target
  form
• If Kevin had studied, he could pass his exam.
• If Kevin would have study, he could have pass the
  exam.

26
How was noticing operationalized?

• Ss in both groups were asked to do the following
• Underline the word, words or parts of words
  that you feel are particularly necessary for your
  subsequent production/reconstruction.
• Rationale online-measure unintrusive quick
  easy
• Since underlining is assumed to involve at least
  a minimum level of awareness, the researchers
  believed that it tapped into detection.

• Just testing!
• Its NOTICING.

27
Procedure

• Both groups were informed of the sequence of
  activities beforehand.
• Treatment for EG Ss had to produce output which
  was followed by exposure to input flood of the
  target form.
• Treatment for CG Instead of producing output, Ss
  answered T/F comprehension questions related to
  the input.
• --------------------------------------------------
  -----------------------------
• Phase 1 Guided Essay-Writing Tasks (A B)
• Phase 2 Text Reconstruction Tasks (A B)

28
ExperimentalSequence
Phase 1 (Task A) 1 hr.
Interview
Phase 1(Task B) 1 hr.
Interview
Posttest 1 (35 mins.)
Phase 2 (Task A) 30 mins.
Interview
Phase 2 (Task B) 30 mins.
Interview
Posttest 2 (35 mins)
29
Phase 1 Guided Essay-Writing Tasks

• EG CG

• EG had to write an essay on a topic that had been
  chosen to elicit the past hypothetical
  conditional.
• Both groups were presented with a model essay on
  the same topic in which 80 of the sentences
  contained the target form.
• EG Produced 2nd version of the essay.
• CG Answered T/F questions based on the model
  essay.

Essay on a given Topic (output 1)
Essay on unrelated topic
Input (reading/underlining)
Essay on the same topic (Output 2)
Comprehension questions
30
Phase 2 Text Reconstruction Tasks

• EG CG

• Read/underlined a text with 70 of the sentences
  containing the past hypothetical conditional.
• EG Reconstructed the passage as accurately as
  possible. (Output 1) CG T/F Comprehension
  Questions
• Read/underlined the same passage as in (Input
  1).
• Same as Output 1 (Output 2)

Input 1 (reading/underlining)
Output 1 (reconstruction)
Comprehension Questions
Input 2 (reading/underlining)
Output 2 (reconstruction)
Comprehension Questions
31
ExperimentalSequence
Phase 1 (Task A) 1 hr.
Interview
Phase 1(Task B) 1 hr.
Interview
Posttest 1 (35 mins.)
Phase 2 (Task A) 30 mins.
Interview
Phase 2 (Task B) 30 mins.
Interview
Posttest 2 (35 mins)
32
Testing Instruments
Multiple-choice recognition test
Example If Jane ________harder last semester, she
__________ her final history exam. a.
studies a. passed
b.will have studies b. had passed
c. had studied c. would
pass d. would study d.
would have passed
Picture-cued production test
Example
B Result
A Option
BECOME
VISIT
If Lisa
33
Scoring Analyses
The corpus of analysis included data from the
participants underlining, written production
during the treatment and the test results.
Underlining
Conditional-related items
Modal Would/Could
Aspectual auxiliaries Have / Had
Copula in the past participle form been
Complementizer If
Past participle endings ed/ en
34
Scoring Analyses
Underlining
the number of conditional-related items
X 100
the total number of items underlined
Production Tasks
the number of correctly formulated target sentence
X 100
the total number of target sentence attempted
Multiple-choice recognition
The recognition test items were scored as either
correct or incorrect.
If-clauses and main clauses were scored
separately, and the score were combined to obtain
a total score for each participant.
35
Scoring Analyses
Production Test
Target-like use analysis
The researchers gave 1 point for each targetlike
production of the conditional form.
Interlanguage analysis
Aimed to examine the specific IL features of
non-targetlike attempts on the production test
and complement the strict coding system used in
the targetlike use analysis.
36
Scoring Analyses
Seven Component Features of the Past
Hypothetical Conditional
If Kentei had traveled
to Korea last summer
past
perfect
past participle
he would have seen
the Word Cup.
past
modal
perfect
past participle
Because of the small sample size and the failure
to obtain a normal distribution, medians were
used as a measure of the central tendency, and
interquartile ranges (IQRs) were used as a
measure of variation in all results reported.
37
Results
Underlining The Noticing Issue
Phase 1
Median Both groups underlined similar median
percentages of conditional-related forms.
The different experimental conditions of the EG
and CG seemed not to contribute significantly to
the extent to which the participants in each
group paid attention to the conditional form.
IQR The IQRs of the EG participants were much
higher than those of the CG.
The EG participants attention was much more
scattered than the CG participants when exposed
to the same input texts.
These results are consistent with those found for
the essay-writing tasks in Izumi et al.s 1999
study.
38
Results
Underlining
Phase 2
Both groups showed the increased percentage of
conditional-related underlining from input 1 to
input 2. No significant differences between two
groups emerged on either input 1 or input 2.
These results show that in both Phase 1 and Phase
2 the output conditions for the EG did not lead
to greater noticing of the conditional form than
did the comprehension-question condition for the
CG.
39
Results
Task Results The Immediate Uptake Issue
Phase 1 Production
1st Essay
2nd Essay
no targetlike use
the correct use of the target form increased
somewhat
These results show that the EG incorporated only
some of the target form in their production
after completing the first guided essay-writing
tasks and being exposed to the input in model
essay.
40
Results
Phase2 Reconstruction
The accuracy scores increased somewhat from the
first to the second reconstruction. This
increase was significant for the combined-clause
scores and for the if- clause score in task A.
In Tasks A and B, the EG participants targetlike
use of the conditional form was much greater
than in essay tasks in Phase 1.
41
Test Results The Acquisition Issue
42
Test Results The Acquisition Issue
43
Test Results The Acquisition Issue
44
Results IL Scoring
The IL scoring seemed to reveal the subtle and
gradual changes in the learners IL. However,
the IL analysis did not show much difference
between the effect of the output and the
comprehension-question treatments.
In all three test administrations, both groups
scored higher on the past feature in the if
clause and on the modal and past features in
the main clause than on the perfect and past
particle features in both clauses.
45
Discussion

• Hypothesis 1
• EG would show greater noticing of the target
  form contained in the input.
• Hypothesis 2
• After being exposed to input containing the
  target form, EG would indicate immediate uptake
  of the target form in their output.
• Hypothesis 3
• EG would show greater gains in accuracy of the
  use of the target form in the posttests.

46
There was no significant between-group
differences on any of the posttest measures.

• Why?
• The common feature of the treatment played a much
  more crucial role than expected.
• What was the common feature of the treatment?
• (Input flood)
• Comprehension task may have inadvertently primed
  the CG to pay closer attention to the target form.

47
Task Effects

• Both Izumi et al.s (1999) study and the current
  study showed that results obtained were
  cumulative and not attributable to just one phase
  of the treatment.
• The EG participants essays showed greater
  individual variation than the text
  reconstructions despite the difference in the
  delivery order of the two types of treatment.
• Essay-writing tasks are susceptible to greater
  individual variation. This type of task
  therefore may not encourage learners to notice a
  specific grammatical form. On the other hand,
  reconstruction tasks that target a specific
  grammatical structure may promote noticing of the
  gap as these tasks maximize the similarities
  between the learners production and learning.

48
Output did not always succeed in drawing the
learners attention to the target form

• This phenomenon seems to be related to individual
  variation and linguistic factors.
• Individual Variation
• After the production task, the EG participants
  varied greatly in their attention to the
  grammatical form while processing the input.

49
Individual Variation

• SA Had some difficulties with the past form while
  writing her first essay, so she paid close
  attention to the past form while reading the
  model essay. (underlined 73)
• When I saw the model, I noticed exactly what
  and where are my mistakes.
• Improvement 6 on pretest to 94 on posttest
• MH His comments almost always focused on the
  ideas or meaning and the organization of his
  essays. (underlined 0- 8)
• Improvement Test scores didnt show measurable
  improvement
• Learners vary greatly in what they find
  problematic in their production and consequently,
  in what they pay attention to in subsequent
  input.

50
Past Hypothetical Conditional
Hypotheticality
Past Time
If Kentei had traveled
to Korea last summer, perfect
past participle past
he would
have seen the Word Cup.
modal perfect
past participle
past
51
IL Scoring
Most participants encoded the features past in
the if clause and modal and past in the
main clause, but they often failed to encode the
features perfect and past perfect in both
if- and main clauses.
The participants major problem with the past
hypothetical form had to do with marking past
time reference rather than marking
hypotheticality.
The participants did not improve on the marking
of the past time reference without also
improving on the marking of hypotheticality.
52
Past Hypothetical Conditional
Formal complexity
Perfective Forms
Have (Had) Past
Participle (aspectual auxiliary)
Functional Expendability
If Kentei traveled to Korea in 1988, he
would see the Olympics in Seoul
Failure to encode the past time reference
features, perfect and past participle,
does not necessarily hinder communication. For
this reason, output or comprehensible input may
not always heighten learners sensitivity to the
past time reference features.
53
Past Hypothetical Conditional
Some SLA literatures (Bardovi-Harlig, 1992, 1999
Meisel 1987 Sato, 1986, Schumann, 1987) point
out that L2 learners have trouble acquiring the
past tense inflection because they rely on the
context or on adverbials such as last night and
yesterday to clarify the past time reference.

A combination of these two reasons formal
complexity and functional expendability- may
have resulted in the participants less noticing
of the past reference features, perfect and
past participle, and consequently they learned
these features lesser degree than they learned
past and modal features.
Like many other pedagogical techniques,
output-input activities may be more effective in
promoting the noticing and learning of some
forms than of others (p. 268)
54
Future Directions

• The Use of Output in L2 Teaching and Learning
• Izume Bigelow suggest using learner output in
  coordination with target language input
• Awareness-raising activities targeted at noticing
  strategies e.g. marking the difference between
  first and revised drafts and reporting on the
  differences noted.
• Output followed by enhanced input
• Completing a reconstruction task collaboratively
  (e.g. dictogloss)

55
Future Directions

• Methodological issues
• What method was used as an online measure of
  learners noticing of the target structure?
• underlining of the input text
• Issues of completeness Does the measure include
  all items that are attended to?
• Issues of precision Does the measure exclude
  items that are not attended to?

56
Future Directions

• Use of multiple measures is important in
  investigating complex notions such as noticing.
  (e.g., online measures, immediate retrospective
  report, task and test rests)
• Further progress in this area of research
  requires a more psycholinguistically
  sophisticated conceptualization of noticing
  before an adequate interpretation of the results
  is possible.

57

Thank You For Your Attention and 'Noticing!'
Have a Wonderful Spring Break!