The effect of increase in task cognitive complexity on Iranian EFL learners’ accuracy and linguistic complexity: A test of Robinson’s Cognition Hypothesis


Payame Noor University, Tehran, Iran


Designing a task with a reasonable level of cognitive complexity has always been important for
syllabus designers, teachers, as well as researchers. This is because task manipulation may lead
to different results in oral production. The present study was an attempt to explore the effect of
this  manipulation  -  based  on  Robinson’s  resource-directing  model  (reasoning  demands,
number of elements, and here and now versus there and then condition) - on picture narration.
The  study  included  30  Iranian  EFL  learners  at  the  intermediate  level  between  the  ages  of  21
and 34. They were all native speakers of Persian. Each participant was required to perform the
simple  version  as  well  as  the  complex  version  of  the  same  picture  narration  task.  The
participants’  speechwas  audio-recorded  and  the  results  revealed  that  an  increase  in  task
cognitive complexity leads to greater accuracy and linguistic complexity.


Main Subjects

For many years, researchers have been eager
to  find  out  whether  different  levels  of  task
structure and cognitive complexity have any
noticeable  impact  on  learners’  oral
production.  This  subject  has  always  been
controversial  since  the  degree  to  adjust  the
task,  according  to  learners’  ability,  has
always attracted the researchers’ attention.  
There is general consensus on the claim that
planning and other factors such as cognitive
complexity  of  the  task  might  have  crucial
impact on oral production (Ahmadian, 2012;
Ahmadian&Tavakoli,  2011;  Ellis,  2000;
Skehan & Foster, 1996; Wendel, 1997; Ellis
Jeon  and  Hahn  (2006)  express  that  task-based  language  teaching  has  a  substantial
implication  for  the  area  of  language
learning.  They  maintain  that  learning  is  a
developmental  process  with  the  aim  of
promoting  communication  and  social
interaction  rather  than  acquiring  a  product
by  practicing  language  items.  Besides,  they
believe  that  learners  learn  the  target
language  more  effectively  when  they  are
naturally  exposed  to  meaningful  task-based
activities.  Although  early  empirical  studies
strongly  support  the  use  of  task  as  a
beneficial  way  to  conceptualize  language
teaching, the amount of research in this area
is still not sufficient. Preparing suitable tasks
demand a great deal of exploration in related
studies  and  a  deep  insight  through  the
influence of the task type on learners’ oral
production  in  terms  of  accuracy  and
linguistic  complexity.  Therefore,  the  use  of
task-based  programs  will  be  open  to  more
research (Skehan, 1998).
According  to  Revesz  (2011),  task
complexity  can  affect  attentional  allocation
and  the  focus  on  second  language  (L2)
constructions  during  task  performance,  and
this  can  influence  the  quality  of  learning.
However,  there  were  some  inconsistencies
among  the  findings  of  applied  linguistics
regarding  the  impact  of  task  type  on
accuracy  and  linguistic  complexity.  To
bridge  the  gap  and  to  understand  the
importance  of  selecting  appropriate  tasks,  it
is  necessary  to  conduct  more  studies  in  the
Literature review    
Since  1980,  second  language  acquisition
(SLA)  researchers  have  been  interested  to
explore  the  impact  of  task  cognitive
complexity  on  oral  production.  A  clear
understanding  of  the  load  of  cognitive
demands  on  participants  can  help  material
producers  to  design  appropriate  tasks  for
learners.  Hence,  tasks’  management  is  a
crucial  basis  for  communicative  language
Skehen’s model of task complexity
To  Skehan  (1998),  three  factors  are
associated  with  task  difficulty:  code
complexity  (the  syntactic  and  lexical
difficulty  of  input),  cognitive  complexity
(the  processing  demands  of  the  tasks),  and
communicative stress (time pressure and the
modality  demand).  These  factors  can
produce  different  demands,  and  therefore
can  influence  the  quality  of  learners’
performance (Taguchi, 2007).
Another factor included in Skehan’s model
is  planning  time.  Previous  research  has
explored  the  effect  of  planning  time  on  L2
output. Planning time would help learners to
produce  more  accurate  as  well  as  greater
level  of  lexical  complexity.  However,  in
case  of  accuracy,  results  could  be  different.
Some studies showed that certain task types
may  lead  to  more  accurate  speech  while
others  proved  that  task  condition  is  an
influential  factor  which  determines  the
degree  of  accuracy.  Skehan  (1998),  as
reiterated in Iwashita, McNamara, and Elder
(2001),  believes  that  different  numbers  of
factors  have  impact  on  task  difficulty.  To
him,  task  dimensions  such  as  abstractness
and  familiarity  of  task  information  can
influence  the  difficulty  of  the  task.  He
maintains  that  performance  conditions  (e.g.,
concrete  vs.  abstract  information)  play  a
crucial  role  in  determining  the  level  of  task
Iwashita,  McNamara,  and  Elder’s  task
According  to  Iwashita  et  al.  (2001),
Skehan’s  framework  was  encouraging;
however,  some  aspects  of  the  framework
were questionable such  as: (1) the notion of
difficulty,  (2)  assessing  the  candidate’s
performance, (3) inconsistent results, and (4)
the  complexity  of  task  performance.  
Iwashita  et  al.  (2001)  consider  four
dimensions  for  narrative  tasks,  with  two
different  performance  conditions  (+  or  -).
These  task  dimensions  and  performance
conditions are defined as follows:
Perspective:  When  a  story  is  told  as  if  it
happened  to  the  participant,  more  accurate
but  less  complex  response  is  produced.
However,  telling  the  story  from  others’
viewpoint  will  make  a  task  more  difficult
and the result would be different.  
Immediacy:If  learners  have  access  to  the
pictures  while  telling  a  story,  their  speech
would  be  more  accurate,  but  less  complex.
On  the  other  hand,  telling  a  story  without
pictures  in  view  would  be  less  accurate;
hence, narrative tasks considering there-and-
then condition are cognitively more complex
than  those  referring  to  here-and-  now
condition.  Iwashita  et  al.  (2001)  associate
this  complexity  with  more  complex  syntax
and multi-propositional utterances.     
Adequacy:  Using  a  complete  set  of  pictures
while  narrating  (-  condition)  would  make  a
task less difficult; in this case learners would
produce  more  accurate,  but  less  complex
sentences.  On  the  other  hand,  if  some
pictures  are  missing,  the  result  would  be
Planning  time:Considering  an  appropriate
planning  time  in  narrative  task  would  make
learners’  oral  production  more  accurate.
However,  narrating  without  considering  the
time limitation may lead to some mistakes in
learners’ oral production.
Robinson’s task complexity dimension
Robinson’s  (2005)  cognition  hypothesis
could  be  the  most  prominent  model  which
was  devised  to  examine  the  impact  of
increase in task cognitive complexity on oral
production.  According  to  Robinson  (2005),
the  cognition  hypothesis  reveals  that  more
difficult  tasks  may  produce  L2  production
with  more  accuracy  and  more  linguistic
complexity.  Based  on  the  hypothesis,
complex  tasks  produce  interactional
processes  such  as  corrective  feedback  and
noticing  of  input.  He  believes  that
researchers  need  to  determine  what
differences L2 tasks, with different levels of
complexity, make to learners’ performance,
so  that  they  can  sequence  and  grade  the
tasks on a proper basis.  
When  learners  perform  more  than  one  task
at the same time, they actually  experience  a
real  world  situation;  in  this  case,  task
complexity  is  increased  along  resource-dispersing  dimensions.  On  the  other  hand,
increasing  task  complexity  along  resource-directing  dimension  (e.g.,  asking  for
justification)  can  motivate  learners  to  use
specific  L2  constructions.  Resource-directing  variables  of  task  complexity
demand  a  great  deal  of  attention  and
working  memory,  and  make  learners  focus
on  linguistic  forms.  Some  examples  of
resource-directing  factors  are:  [±  few
elements],  [±here  and  now],  and
[±reasoning].  The  low  complexity
conditions include [+ few elements], [+ here
and  now],  and  [-  reasoning]  and  the  high
complexity conditions are associated with [-
few  element],  [-here  and  now],  and
[+reasoning] (Robinson, 2001; 2005).  
According  to  Robinson,  learners’  attention
can  be  diverted  over  many  L2  elements
when the task complexity is increased along
recourse-dispersing  dimensions.  Some
examples  of  resource-dispersing  factors
include:  [±  planning], [±single  task],  and  [±
prior  knowledge].  The  low  complexity
conditions  are  [+planning],  [+single  task],
and  [+prior  knowledge]  while  the  high
complexity  conditions  include  [-planning],
[-single  task],  and  [-prior  knowledge]
(Robinson,  2005).  The  cognition  hypothesis
proves  that  when  we  increase  the  cognitive
complexity  of  the  task,  learners  show  more
accurate,  but  less  fluent  language.
Furthermore,  in  more  complex  tasks,
interactional  processes  such  as  noticing  and
corrective  feedback  are  noticeable  (Kim,
This  study  was  an  attempt  to  explore  the
extent  to  which  cognitive  complexity  in
tasks  could  have  an  effect  on  Iranian  EFL
learners’  oral  production.  To  identify  the
relations,  contradictions,  and  gaps  in  the
literature,  the  following  research  questions
were  formulated  to  check  the  aim  of  the
1.  Does  increase  in  task  cognitive
complexity  affect  the  accuracy  of
Iranian  EFL  learners’  oral
2.  Does  increase  in  task  cognitive
complexity  affect  the  linguistic
complexity of Iranian EFL learners’
oral production?     
This  quasi-experimental  research  drew
preliminary  on  Robinson’s  cognition
hypothesis  which  was  a  foundation  for
investigating  the  impact  of  cognitive
complexity  on  the  aspects  of  oral
performance such as accuracy and linguistic
The  participants  in  this  study  were  30
students  at  the  intermediate  level  between
the  ages  of  21  to  34.  They  were  all  female
students  studying  general  English  at  Kish
Institute  in  Tehran,  Iran.  The  participants
were  all  native  speakers  of  Persian,  and  on
average  they  had  been  studying  English  for
three years. Based on non-random sampling,
the  participants  were  selected  from  two
intact  classes,  with  15  participants  in  each.
A  version  of  Test  of  English  as  a  Foreign
Language  (TOEFL)  was  administrated  to
assure the homogeneity of the participants in
the study.
Testing materials
In  order  to  evaluate  learners’  oral
production,  two  tests  were  used  by  the
researchers:  speaking  tests  from  TOEFL
IBT  book  by  McGraw  (2006)  for  checking
the  homogeneity  of  the  participants,  and  a
post-test  with  the  aim  of  measuring  and
comparing oral skill of the two groups.  
Teaching materials
For  teaching  materials  and  treatment,  some
narrative  tasks  were  selected  from  English
Result  by  Hancock  and  McDonald  (2012)
with  tasks  and  the  exercises  designed  for
intermediate  level.  The  teacher  and  the
researchers agreed that the tasks in this book
were appropriate for the aim of the treatment
because  the  content  of  the  book  covered
both simple as well as complex tasks which
could be suitable for assessing learners’ oral
skills  according  to  resource-directing
Data collection and procedure
The data in this study drew mostly from the
participants’  oral  production  which  was  in
the  form  of  picture  narration  tasks.  All  the
necessary data were collected during one of
the students’ regular term. The classes lasted
two  hours,  and  two  experimental  groups
participated in this study. For the aim of the
study,  three  elements  of  Robinson’s
resource-directing  model,  +/-  few  elements,
+/- reasoning demand, and +/- here and now
condition, were checked along accuracy and
linguistic complexity.  
To  ensure  that  the  tasks’  design
manipulation  was  appropriate  for  the
purpose of the study, two raters, experienced
English  teachers,  cooperated  with  the
researchers  in  estimating  the  level  of
cognitive  complexity  of  the  tasks.  For  each
element of this model, the participants were
asked  to  perform  two  versions  of  the  same
task  (+  condition  and  –  condition).  Hence,
six sources of data were analyzed to answer
the two research questions.
Checking homogeneity
Prior  to  the  administration  of  the  tasks,  to
assure  homogeneity  of  participants,  some
speaking  tests  from  TOEFL  IBT  book  by
McGraw  (2006)  were  chosen  and  the
participants’ voices were recorded; based on
Speaking  Rubric  the  responses  were  scored
from 4 to 0. The results of the independent t-test  [t  (28)  =  .843,  p  =  .406]  indicated  that
the  two  groups  were  homogeneous  in  terms
of  their  general  language  proficiency.  The
mean  scores  for  the  two  groups  were  17.66
and  17.31  respectively  which  point  to
homogeneity of the two groups.
Treatment.  The  treatment  plan  for  the  first
experimental  group  included  working  on
more  difficult  tasks  (see  Appendix  A).
These  tasks  demanded  more  causal
reasoning  as  well  as  justification  for  the
replies.  The  practice  plan  for  these
participants  included  all  the  aspects  which
were  aimed  at  the  post-test  such  as  +/-
reasoning  demand,  +/-  few  elements,  and
here and now vs. there and then condition.   
For the reasoning demand aspect of the task,
two  sets  of  pictures  were  selected,  one  with
correct  order  and  the  other  with  scrambled
pictures. For checking the impact of number
of elements on learners’ oral production, the
researchers  asked  the  participants  to  narrate
the  story  once  with  9  pictures  and  the  other
time  with  6  pictures  (the  three  last  pictures
were omitted).  
For  the  last  aspect  of  Robinson’s  (2001)
resource-directing  model,  here  and  now  vs.
there  and  then  condition,  the  participants
were  first  required  to  tell  the  story  with
pictures  in  front  of  them.  Next,  the
participants  were  asked  to  turn  the  picture
strips over before beginning their narration.  
For  both  narrations,  the  participants  were
given  prompts  and  instruction.  For  each
dimension,  two  conditions  (+  condition  and
– condition) were needed to be tested, so all
the participants in both experimental groups
were supposed to Perform six tasks. Table 2
illustrates  a  brief  description  of  test  tasks
based  on  Robinson’s  model.  The
participants  in  the  second  experimental
group  were  exposed  to  simpler  tasks  in  the
form  of  picture  narration  (see  Appendix  B).
For  the  simpler  task,  a  set  of  four  pictures
were  selected  with  a  topic  familiar  to  the
participants,  and  the  task  did  not  require
causal  reasoning,  justification  of  beliefs,  or
any kind of interpretation.  
Post  Test.  The  post-test  was  administered
one  week  after  the  treatment.  The
researchers  preferred  a  monologic  picture
description  task  to  elicit  participants’  oral
performance (see Appendix C).  
As mentioned earlier, each group performed
two  versions  of  the  same  narrative  task:  a
simple  and  a  complex  version.  The
participants  had  three  minutes  to  read  the
instructions,  take  note,  and  prepare  their
Therefore,  the  three  mentioned  dimensions
in  Robinson’s resource-directing  model,  [±
few  elements],  [±here  and  now],  and
[±reasoning],  were  examined  in  the  study.
As  Table  1  shows,  the  low  complexity
conditions include [+ few elements], [+ here
and  now],  and  [-  reasoning]  while  the  high
complexity conditions are [- few element], [-here and now], and [+reasoning].  
It  is  also  worth  mentioning  that  in  order  to
avoid  practice  effect,  counterbalancing  was
suggested by the researchers.

Rating  Scale. To measure participants’ oral
production  in  terms  of  accuracy  and
linguistic  complexity  in  the  post-test,  an
analytic rating scale proposed by Iwashita et
al.  (2001)  was  chosen.  In  this  rating  scale,
linguistic control as well as managing forms
and grammar are among the most prominent
factors  while  assessing  accuracy.  Based  on  
Iwashita  et  al.  (2001),  attempting  a  variety
of  verb  forms  (e.g.,  passive,  modals,  and
tense),  taking  grammatical  risks  in  order  to
express  complex  meaning,  and  using
coordination  and  subordination  to  transfer
ideas were the basis for checking linguistic  

The  results  of  MANOVA  are  presented  in
two  subcategories:  The  impact  of  the
increase  in  task  cognitive  complexity  on
accuracy and linguistic complexity.
The impact of the increase in task  cognitive
complexity on accuracy
This  part  depicts  the  analysis  and  findings
for the first research question: Does increase
in  task  cognitive  complexity  affect  the
accuracy of learners’ oral production?
Based on the outcome of MANOVA, the two
experimental groups’ means were compared
on  the  three  accuracy  tests  (the  elements  of
Robinson’s  model).  According  to  Table  3,
the following results were obtained:
F  (3,  36)  =  11.24,  p  =  .00,  and  Partial  η2  =
As  the  Table  reveals,  the  F-observed  value
calculated  for  the  effect  of  the  difficulty
level  on  the  students’  overall  accuracy  in
oral production was statistically significant.

The statistics on the mean score displayed in
Table 4 and Table 5 revealed that on average
the  experimental  group  1  (M  =  1.93,  SD  =
.65)  was  more  successful  than  the
experimental  group  2  (M  =  1.33,  SD  =  .57)
on the accuracy in tasks checking reasoning
demands  ,  F  (1,  28)  =  12.06,  p  =  .002,  and
Partial η2 = .30.

Since  the  F-value  of  11.24  indicated  a    
significant  difference  between  the
experimental  group  1  (with  difficult  tasks)
and  the  experimental  group  2  (with  easy
tasks), the two groups’ performances on the
three  tasks  including  reasoning  demands,
number  of  elements,  and  here-and-now  vs.
there-and-then condition were compared.
The  results  displayed  that  the  experimental
group 1 (M=1.93, SD=.50) had higher means
than the experimental group 2 (M = 1.46, SD
=  .69)  on  the  accuracy  tests  while  checking
number  of  elements.  As  shown  in  Table  5,
there  was  a  significant  difference  between
the two groups’ means on the accuracy tests
as  far  as  the  number  of  elements  was
concerned,  [F  (1,  28)  =  6.86,  p=  .00,  and
Partial η2 = .19].   
Accordingly, another MANOVA test was run
to check the results of groups’ performance
in  tasks  considering  accuracy  with  two
conditions  (here-and-now  vs.  there-and-then). A simple comparison, based on Table
4, demonstrated that the experimental group
1  (M  =  1.93,  SD  =  .47)  showed  greater
efficacy than the experimental group 2 (M =
1.33,  SD  =  .55)  on  the  accuracy  in  tasks
checking  the  mentioned  conditions.  
Furthermore,  according  to  Table  5,  there
was a large effect size [F (1, 28) = 12.06, p
= .002, and Partial η2 = .30]. In other words,
there  was  a  significant  difference  between
the  two  groups’  performance  on  the
accuracy of here-and-now vs. there-and-then

As  Figure  1  and  the  obtained  information
reveal,  the  first  null  hypothesis  could  be
The impact of the increase in task cognitive
complexity on linguistic complexity
In  this  section,  we  investigate  the  findings
for  the  second  research  question:  Does
increase  in  task  cognitive  complexity  affect
the  linguistic  complexity  of  learners’  oral
The results of performing another MANOVA
test  indicated  that  the  F-observed  value
obtained from the students’ performances on
overall  linguistic  complexity  was
statistically  significant  [F  (3,  36)  =  24.68,
p=  .00,  and  partial  η
  =  .74].  The  analysis
also  specified  a  large  effect  size  (see  Table
6);  hence,  the  second  null-hypothesis  could
be rejected

According  to  Table  6,  the  F-value  of  24.68
revealed a significant difference between the
means  on  linguistic  complexity  of  oral
production;  however,  the  two  groups’
performance  on  the  three  elements  of
Robinson’s  model  was  also  compared.
Based on the descriptive statistics displayed
in  Table  7  and  Table  8,  it  could  be  realized
that on average the experimental group 1 (M
= 2.06, SD = .60) had better results than the
experimental  group  2  (M  =  1.13,  SD  =  .60)
on  the  linguistic  complexity  in  reasoning
demands  tasks.  Comparing  the  results  in
Table 8 proved a large effect size revealing a
significant  difference  between  the  two
groups’  means  on  linguistic  complexity  [F
(1,  28)  =  27.44,  P  =  .00,  and  Partial  η2 = .49].
The  data  in  Table  7  depicted  that  the
experimental  group  1  (M  =  2.20,  SD  =  .69)
outperformed the experimental group 2 (M =
1.06,  SD=  .68)  on  linguistic  complexity
while  performing  tasks  with  different
number  of  elements.  According  to  Table  8,
there  was  a  large  effect  size  [F  (1,  28)  =
36.78,  p  =  .00,  and  Partial  η2 =  .56]. the
data proved a significant difference between
the  two  groups’  means  on  linguistic
complexity in tasks with different number of

As  displayed  in  Table  7,  on  average,  the
experimental  group  1  (M  =  2.06,  SD  =  .74)
got  better  scores  than  the  experimental
group  2  (M  =  1.53,  SD  =  .49)  on  linguistic
complexity  on  here-  and-now  vs.  there-and-then  condition.  The  obtained  data  from
Table 8, [F (1, 28) = 6.89, p = .01], reveals a
significant  difference  between  the  two
groups’  means  on  linguistic  complexity
while  performing  the  tasks  in  two
conditions.  Figure  2  supports  the  statistical

In  the  case  of  oral  production,  adjusting  the
complexity  of  the  task  appears  to  be  one  of
the  prominent  aims  of  syllabus  designers
because  based  on  different  studies  such  as
the  ones  conducted  by  Revesz  (2011),
Iwashita, et al. (2001), and Robinson (2005),
human  beings  possess  a  limited  processing
capacity  and  are  not  capable  of  attending
fully to all aspects of a task. When different
aspects of oral production are concerned, an
appropriate level of  cognitive complexity in
tasks  would  help  learners  to  promote  their
speaking.  This  study  sought  to  explore  the
effects  of  increase  in  task  cognitive
complexity  on  accuracy  and  linguistic
complexity  of  Iranian  EFL  learners’  oral
production in narrative tasks.  
In general, the results of the study indicated
that  the  group  that  had  exposure  to  more
complex  tasks  presented  more  accurate
speech  with  more  complex  structure.  This
difference  in  the  outcome  was  noticeable  in
all the three aspects of Robinson’s resource-directing model.
The  effect  of  increase  in  task  cognitive
complexity on accuracy
The  findings  revealed  that  scrambled
pictures,  designed  for  checking  reasoning
demands,  resulted  in  more  accurate  speech.
The  major  explanation  for  this  difference
could  be  that  while  telling  a  story  with
scrambled pictures, participants tried to rely
on  their  logical  concepts  in  order  to  put  the
pictures  in  the  correct  order  on  their  own
way,  and  for  doing  that  they  needed  to  get
help from their knowledge repertoire.  
The  results  concerning  accuracy  are  in  line
with  Skehan’s  (1998,  as  cited  in  Taguchi,
2007)  study.  The  findings  also  support  the
claim  of  Iwashita  et  al.  (2001)  about  the
effect  of  complex  tasks  on  directing
learners’  attentional  resources  to  forms
which can lead them to induce risk avoiding
behavior. Robinson (2001; 2005, as  cited in
Revesz,  2011)  also  emphasizes  the
beneficial  role  of  complex  tasks  on  making
learners  focus  on  linguistic  forms.  Besides,
Tavakoli’s  (2009)  cognitive  approach
towards  language  learning  supports  the
findings in this study.  
In the case of number of elements, the group
that  performed  the  narration  task  with  more
pictures outperformed the group that did the
task  with  fewer  pictures.  This  might  be
because using more pictures while narrating
produces  more  sentences,  so  participants  
were  concerned  with  the  correct  connection
in  order  to  make  their  sentences
comprehensive.  This  made  the  participants
process  their  speech  in  their  minds  before
revealing  it.  This  reformulation  and
consistent  monitoring  resulted  in  more
accurate  speech.  The  outcome  bears  out  the
previous  studies  conducted  by  Kim  (2009),
Iwashita  et  al.  (2001),  and  Robinson  (2001,
2005, as cited in Revesz, 2011).  
The  study  also  highlights  the  findings
regarding  accuracy  while  checking  here-and-now vs. there-and-then condition. In the
case  of  telling  a  story  with  pictures  in  view
(here-and-now),  the  participants  produced
less  accurate  sentences  than  when  they  had
performed the same task without pictures in
front  of  them  (there-and-then).  The  point  is
that  there-and-then  condition  (telling  the
story without pictures) put more burdens on
participants’  working  memory  and  the
participants paid more attention to the forms
before  producing  their  speech,  therefore
their  speech  were  more  accurate.  This
viewpoint  is  in  accordance  with  Revesz’s
(2011) statement.  
The  effect  of  increase  in  task  cognitive
complexity on linguistic complexity
The  results  of  the  study  revealed  that
increasing  the  level  of  task  difficulty  led  to
more  complex  speech  in  all  the  three
elements  of  Robinson’s  model.  In  fact,  in
telling  a  story  with  scrambled  pictures,
participants  produced  more  complex
sentences  than  the  time  they  were  exposed

to  pictures  in  the  correct  order.  Scrambled
pictures  made  participants  process  the
information in their minds while seeking for
some  new  and  more  complex  structures.
Besides,  participants  were  more  willing  to
provide  reasons  to  prove  why  they  put  the
pictures  in  that  order.  This  led  them  to  use
new  and  more  complex  structures;  this
stance  is  supported  by  Robinson’s  (2001,
2005, as cited in Revesz, 2011) model.  
The findings demonstrated that narrating the
story  with  more  pictures  led  to  more
complex language. This  could be due to the
use  of  more  complex  syntax  which  was
associated with the tendency of using more
transitions  and  coordination  conjunctions.
Other scholars such as Talmy (2000, as cited
in  Revesz,  2011)  pointed  to  this  impact  as
As far as here-and-now vs. there-and-then is
concerned,  it  could  be  realized  that  in
second  condition  participants  tended  to
remember the events and they relied on their
memory  to  connect  the  sentences  in  a
coherent way and that led the participants to
produce  not  only  more  accurate  speech  but
also  more  complex  syntax.  This  attitude
supports Iwashita et al.’s (2001) approach in
that more attention to forms and planning of
production is due to the greater demands on
memory  and  the  attempt  to  make  transition
between the events.
Generally,  Iranian  EFL  learners  have
difficulty  producing  an  accurate  speech.  To
solve  this  problem,  we  need  a
comprehensive  insight  into  methodology  in
general  and  designing  and  adjusting  the
level  of  tasks  in  particular.  Based  on  the
findings of the present study, the  group that
performed  more  cognitively  difficult  tasks
was  more  successful  in  terms  of  accuracy
and linguistic complexity; therefore, the first
and  second  null  hypotheses  were  rejected.
The results confirm the beneficial impact of
increasing  the  task  cognitive  complexity  on
speaking,  especially  in  accuracy  and
linguistic complexity. This in turn can be an
acceptance  for  Robinson’s  cognition
hypothesis  with  his  emphasis  on  the
promotion of learners’ oral skills along with
opting  for  more  challenging  tasks.  This
notion  and  the  previous  studies  as  well  as
the  results  of  the  current  research  indicate
that increasing the difficulty of the task to a
reasonable  level  can  be  highly  effective  for
learners’ improvement in speaking.  
However, there were some limitations in the
current study. First, gender differences were
not  considered  in  this  research;  the
participants  were  all  females,  hence  the
results  may  not  be  generalized  to
coeducational systems. Second, choosing the
non-random  sampling  method  and  selecting
the participants from intact classes as well as
small  sample  size  could  have  influence  on
external validity.

The  present  study  has  been  supported  by
Payame Noor University, Iran.


Ahmadian,  M.  J.  (2012).  The  effects  of
guided  careful  online  planning  on
complexity, accuracy, and fluency in
intermediate  EFL  learners’  oral
production:  The  case  of  English
articles.  Language  Teaching
Research,  16,  127–
Ahmadian, M. J., &Tavakoli, M. (2011).The
effects of simultaneous use of careful
online  planning  and  task  repetition
on accuracy, fluency, and complexity
in  EFL  learners’  oral  production.
Language  Teaching  Research,  15,
Ellis,  R.  (2000).  Task-based  research  and
language  pedagogy.Language
Teaching Research, 4(3), 193-220.
Ellis,  R.  (2003).  Task-based  language
learning  and  teaching.  Oxford:
Oxford University Press.
Hancock,  M.,  &  McDonald,  N.  (2012).
English  result.    Oxford:  Oxford
University Press.
Iwashita,  N.,  McNamara,  T.,  &    Elder,  C.  
(2001).  Can  we  predict  task
difficulty in an oral proficiency test?
Exploring  the  potential  of  an
information-  processing  approach  to
task  design.Language  Learning,
51(3):  401-436.
Jeon,  I.  &  Hahn,  J.  (2006).Exploring  EFL
teachers’  perception  of  task-based
language  teaching:  A  case  study  of
Korean  secondary  school  classroom
practice.Asian  EFL  Journal,  8
(1).Retrieved  May,  15,  2013  from
Kim,  Y.  (2009).  The  effect  of  task
complexity  on  learner-learner
interaction.System, 37: 254-268.
McGraw, H. (2006). The official guide to the
new  TOEFL  IBT.  New  York,  NY:
Revesz,  A.  (2011).  Task  complexity,  focus
on  L2  construction,  and  individual
differences:  A  classroom-  based
study.  Modern  Language  Journal,
95: 162-181.
Robinson,  P.  (2001).  Task  complexity,  task
difficulty,  and  task  production:
exploring  interactions  in  a
componential  framework.  Applied
Linguistics,  22,  27–57.doi:  
Robinson,  P.  (2005).  Cognitive  complexity
and  task  sequencing:  Studies  in  a
componential  framework  for  second
language  task  design.  International
Review of Applied Linguistics, 43: 1-32.
Skehan,  P.  (1998).  A  cognitive  approach  to
language  learning.Oxford:  Oxford
University Press.
Skehan, P. & Foster.P. (1996).The influence
of  planning  on  performance  in  task-based  learning.Studies  in  Second
Language Acquisition, 18, 299-324.
Taguchi,  P.  (2007).  Task  difficulty  in  oral
speech  act  production.Applied
Linguistics, 28(1), 113-135.
Tavakoli, P. (2009). Task design and second
language  performance:  The  effect  of
narrative    type  on  learner  output.
Language Learning, 56(2), 439-473.
Wendel,  J.  N.  (1997).  Planning  and  second
language  narrative
production.Unpublished  doctoral
dissertation,  Temple  University,