-
Applied Linguistics 28/1: 113135 Oxford University Press
2007doi:10.1093/applin/aml051
Task Difficulty in Oral Speech ActProduction
NAOKO TAGUCHI
Carnegie Mellon University
This study took a pragmatic approach to examining the effects of
task difficulty
on L2 oral output. Twenty native English speakers and 59
Japanese students
of English at two different proficiency levels produced speech
acts of requests
and refusals in a role play task. The task had two situation
types based on three
social variables: interlocutors power difference (P), social
distance (D), and the
degree of imposition (R). In one situation type, the power
relationship was
equal, the distance was small, and the degree of imposition was
small
(PDR-low). In the other situation type, the listener had greater
power, the
distance was large, and the degree of imposition was also large
(PDR-high).
The participants production was analyzed for overall
appropriateness (rated on
a 6-point scale), planning time, and speech rate. Results showed
that L2 learners
produced PDR-low speech acts significantly more easily and
quickly, but little
difference was observed in native speakers production. There was
a significant
proficiency effect on appropriateness ratings and speech rate,
but not on
planning time. Post hoc analyses showed that each group
demonstrated different
patterns in the choice of linguistic expressions over the two
situation types,
indicating the noteworthy impact of situational variation on
oral speech
act production.
In the field of second language acquisition (SLA), there has
been growing
attention on the nature of tasks in which second language (L2)
learners
engage to achieve language functions (Bygate et al. 2001;
Crookes and Gass
1993; Ellis 2003; Robinson 2001; Skehan 1996, 1998). When
describing
tasks, previous research mainly used variables from a cognitive,
information-
processing perspective to operationalize task difficulty (e.g.
the demands on
memory, time pressure) (Skehan 1996, 1998). For instance, tasks
that use
unfamiliar information, involve numerous steps for completion,
and provide
no planning time are considered more difficult to perform than
simpler,
familiar tasks that involve only a few operations and provide
plenty of
planning time. These psycholinguistic dimensions have been
manipulated in
developing different task types. Previous research has analyzed
learner
production elicited through different tasks or the same task
with different
variables, and provided empirical evidence that features of L2
oral output,
such as accuracy, fluency, and complexity, vary by task type
(e.g. Bygate
1999; Chang 1999; Robinson 1995, 2001; Robinson et al. 1995;
Skehan and
Foster 1999; Wigglesworth 1997, 2001).
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
Although previous findings have generally confirmed that the
psycho-
linguistic dimensions of tasks have consistent effects on
performance, what
is lacking in the existing literature is the inclusion of
pragmatic conditions in
defining task difficulty. Because the central objective of
language teaching is
to develop learners communicative abilitythe ability to use
linguistic
resources to perform functions in contextrecent literature has
focused on
the social dimensions of language tasks. Within L2 pragmatics in
particular,
social and interpersonal factors (e.g. speakers social distance
and the degree
of imposition), have been incorporated into tasks to elicit L2
pragmatic
production. Following this trend, this study takes a pragmatic
approach
to examining task difficulty. It uses social and interpersonal
dimensions to
develop task types. Learner output is analyzed for
appropriateness and
production speed to examine the kind of impact such task
variation might
have on L2 oral output. This investigation intends to provide
insights into
what makes a task more difficult pragmatically and how the task
difficulty
is manifested in learners oral production.
BACKGROUND
A task is described as a goal-oriented activity involving a
meaningful,
real-world process of language use, and engages four language
skills as well
as cognitive processes (Ellis 2003). A main objective in
researching language
tasks has been to identify a set of task characteristics based
on the assumption
that learner performance varies according to task
characteristics. Recent
research has examined how different task types affect the
quality of learner
output in classroom settings and in the context of assessment
(Bygate 1999;
Bygate et al. 2001; Candlin 1987; Long and Crookes 1992; Nunan
1989;
McNamara 1996; Norris 2002; Norris et al. 1998; Robinson 2001;
Skehan
2001; Swain and Lapkin 2001). The mainstream approach has been
to apply
psycholinguistic categories to operationalize the degree of task
difficulty
(Robinson 2001; Skehan 1996, 1998). Drawing on Candlins (1987)
claim,
Skehan (1996, 1998) outlined three factors that contribute to
task difficulty:
code complexity (the syntactic and lexical difficulty of
language input),
cognitive complexity (the processing demands of tasks such as
memory and
attention) and communicative stress (e.g. time pressure and the
modality
demand). These psycholinguistics factors, once incorporated into
the task
structure, are considered to produce differential demands, and
consequently
affect the quality of learners performance (Robinson 2001).
Previous researchers manipulated these psycholinguistic factors
to design
tasks to elicit learner output, and the output was analyzed with
the measures
of accuracy, fluency, and complexity of language (Chang 1999;
Robinson
1995, 2001; Robinson et al. 1995; Skehan and Foster 1999;
Wigglesworth
1997, 2001). In Robinsons (2001) study, 44 Japanese learners of
English
completed two map tasks: a simple task involving giving
directions to places
in a small, familiar area and a complex task involving giving
directions to
114 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
locations within a large, unfamiliar area. It turned out that
neither task had
significant impact on syntactic complexity, as measured by the
number of
subordinate clauses produced, but in the complex task, the
production was
more accurate (as measured by the number of error-free clauses)
and less
fluent (as measured by the number of words per clause).
Another psycholinguistic factor in task difficulty, planning
time (a factor
included in Skehans model of communicative stress) has been the
focus of
a number of studies, based on the assumption that the quality of
production
depends in part on efficient planning. Previous research has
examined the
effect of planning time on L2 output, yielding generally
confirmed findings
that planning time positively influences linguistic output
(Crookes 1989;
Ellis 1987; Foster and Skehan 1996; Mehnert 1998; Ortega 1995;
Skehan and
Foster 1997; Wigglesworth 1997; Yuan and Ellis 2003).
Wigglesworth (1997)
examined the effect of planning time on 107 learners of English
in spoken
test. Results revealed that, for high proficiency learners,
planning time
helped to produce complex, fluent language, but there was no
effect on
accuracy. Yuan and Ellis (2003) examined the effects of pre-task
and online
planning on speech. Forty-two learners of English narrated a
story orally
based on a picture. The learners were divided into three task
conditions:
no-planning, pre-task planning in which they received 10 minutes
to plan,
and online planning in which they had unlimited time to narrate
the story.
Results showed that pre-task planning enhanced grammatical
complexity,
fluency, and lexical variety, while online planning promoted
accuracy.
To summarize, previous research operationalized task difficulty
mainly
from a psycholinguistic perspective, incorporating factors that
are likely to
affect learners cognitive processing load (e.g. time pressure,
topic
familiarity). The findings showed that tasks with greater
cognitive demands
affect L2 spoken output in terms of accuracy, fluency, and
complexity.
However, questions remain as to whether the concept of task
difficulty can
be applied to other task types, specifically, those tasks that
are more
pragmatic-oriented in design and thus reflect social and
interpersonal
conditions. It is questionable whether the social and
interpersonal
dimensions of a task, if manipulated to increase or reduce task
demands,
could influence learner output in measurable ways, and whether
the effect is
observed differently according to learners proficiency levels.
Because
different tasks are likely to encourage different types of
language processing,
tasks that are pragmatic-orientated should be included in the
line of research
on task difficulty.
Such investigation will be useful because it corresponds to the
current
emphasis on the functional, communicative use of language in
context in
L2 instruction (e.g. Council of Europe 2001). Paralleling this
trend,
developmental studies of pragmatic competence have received
much
attention (Kasper and Rose 2002). Investigation of pragmatic
task
characteristics could be a way to examine pragmatic development,
because
it informs us which pragmatic tasks are more difficult to
perform and thus
NAOKO TAGUCHI 115
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
are mastered at a later stage of development, as well as what
factors create
such task demands. Thus, in addition to the psycholinguistic
approach,
a pragmatic approach should be applied to define task
difficulty.
When developing pragmatic tasks, previous research used a unit
of speech
acts (e.g. requests) following Speech Act Theory (Austin 1962;
Searle 1969).
According to Brown and Levinson (1987), social and interpersonal
factors,
namely interlocutors power difference, the social distance
between them,
and the degree of imposition, influence the directness levels of
speech act
expressions. In a situation where the speech act involves a high
degree of
imposition, is addressed to a person who has more power, and is
in a more
distant relationship (e.g. asking a teacher for a recommendation
letter),
a greater degree of politeness is required to allow the
interlocutor to save
face. In contrast, when the speech act involves a low degree of
imposition
and is produced for a person in an equal relationship (e.g.
asking a friend for
a pen), a lesser degree of politeness is required. Thus, the
social factors of
power, distance, and imposition are thought to make speech acts
more
demanding to perform in certain situations than in others.
These social factors have been included in L2 speech act
research to
develop different task situations (e.g. Blum-Kulka et al. 1989;
Beebe et al.
1990; Maeshiba et al. 1996; Rose 2000; Trosborg 1995). These
studies have
examined learners sociolinguistic sensitivity: whether learners
could vary
their linguistic forms according to situations, thereby
approximating native
speaker norms. Findings showed that the social factors in tasks
affected
linguistic expressions, and the expressions differed according
to learners
proficiency. Advanced learners used more native-like expressions
across
situations.
Although abundant research has examined L2 speech act production
over
different social situations, little has been done to explicitly
link the concept
of task difficulty to the social demands of a task, as well as
to the nature of
L2 output affected by the demands. Fulcher and Marquez-Reiters
(2003)
study is one of the few in this direction. Their study examined
a relationship
between social features of speech act tasks and the perceived
difficulty of the
tasks by native speakers. Twenty-three Spanish- and 32
English-speaking
students watched video recordings of their role play performance
of requests
and judged how successful their requests were, based on a
10-point scale.
Their perceived degree of task difficulty was notably high in
the extreme
cases of social power and imposition. Fulcher and
Marquez-Reiters study
suggests that social factors such as power, relationship, and
the degree of
imposition could serve as useful factors in creating contrasting
task conditions
and predict task difficulty, consequently affecting output in
measurable ways.
More empirical effort is needed in L2 to confirm the usefulness
of the social
factors in conceptualizing task difficulty.
When conceptualizing task difficulty, measures of language
output elicited
through a task require careful consideration because they reveal
how
features of learner output vary according to tasks. Previous
research used the
116 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
measures of fluency, accuracy, and complexity of linguistic
output. In speech
acts, however, specific measures should be adapted to examine
the impact
of social demands of tasks on learner output. Previous studies
mainly
measured L2 speech acts based on learners knowledge, as
exemplified in the
types of linguistic expressions, and the extent to which they
approximated
native speaker choices (e.g. Maeshiba et al. 1996; Rose 2000;
Trosborg 1995).
Other features that could serve as measures, such as overall
appropriateness,
planning time, and oral fluency, have usually been neglected,
consequently
limiting the analysis to syntactic features. As Kasper (2001)
states, pragmatic
competence refers to the acquisition of pragmatic knowledge and
to gaining
automatic control in processing it in real time. Thus,
indications of
performance fluency, such as planning speed and oral fluency,
which have
been used in task-related research, could provide additional
useful
information from a processing perspective, that is, how rapidly
they can
process pragmatic knowledge and convey speech intentions.
The knowledge dimension of speech acts is reflected in learners
ability
to use appropriate speech act expressions. Pragmatic knowledge
of appropri-
ateness reflects two major concepts: sociopragmatics (i.e.
evaluation
of contextual factors) and pragmalinguistics (i.e. linguistic
resources available
to perform language functions) (Kasper 1992; Leech 1983; Thomas
1983).
As Leech (1983) argues, pragmalinguistics is applied to the
study of the more
linguistic and grammatical end of pragmatics, while
sociopragmatics
is applied more toward the sociocultural end. Both are necessary
elements
of pragmatic knowledge that determine appropriateness of L2
speech act
production.
The processing dimension of speech acts, on the other hand, is
reflected
in learners cognitive efficiency in accessing and processing
pragmatic
knowledge. Speed of performance is considered an indirect
reflection of such
cognitive efficiency and provides an indication of processing
abilities, rather
than knowledge (Juffs 2001; Lennon 1990). According to Levelt
(1989),
speech production involves three interacting components:
conceptualizer,
formulator, and articulator. Conceptualizer is the
message-generation phase
where planning of the discourse direction occurs. Pre-verbal
message then
moves to the formulation stage, where lemmas are activated,
and
grammatical and phonological encoding take place. The encoded
message
then becomes articulated. These different components work
simultaneously.
De Bot (1992) adapted Levelts model to bilingual speakers.
According
to de Bot, the conceptualizer is partly language-specific and
partly
language-independent. Different formulators exist for each
language, while
there is one lexicon. The articulator uses language independent
speech
motor plans.
As illustrated in the models, speaking is a process of
translating concepts
into linguistic units and articulating them online. It involves
the process
of planning what to say and how to say it (Foss and Hakes,
1978;
Wiese 1985). Speakers evaluate linguistic and non-linguistic
knowledge
NAOKO TAGUCHI 117
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
and make decisions about how to convey communicative intentions,
and
articulate intentions using linguistic means. Fluent speech acts
result from
rapid transformation of social concepts and articulation of the
concepts.
In speech acts non-linguistic, social knowledge, such as speaker
relationship
and goals of interaction, contributes to planning. Thus, two
dimensions of
production, planning speed and speaking speed, could jointly
promote fluent
speech acts.
In summary, with increasing attention being given to the
characteristics
of language tasks in communicative situations, alternative means
to
operationalize task characteristics are necessary. One means is
to extend
the notion of task difficulty to speech act pragmatics by
studying the social
dimensions of tasks and their impact on oral speech act
production. When
examining the impact of task demands, measures of processing
capacity, such
as speed in planning and articulation, combined with the
measures
of appropriateness, could provide more complete information on
the quality of
L2 speech acts across different task situations. These multiple
criteria could
provide a means to measure the impact of task types on L2
output, and thereby
enhance our understanding of task difficulty in pragmatic
production.
RESEARCH QUESTIONS
This study used social variables to develop different task
situations for speech
acts and analyzed L2 speech acts according to three criteria:
overall
appropriateness, planning time, and oral fluency operationalized
as speech
rate. The study also examined whether the task characteristics
showed
differential effect according to learners proficiency levels.
Three research
questions guided this investigation:
1 Do the type of social situations and L2 proficiency have
differential effects
on the appropriateness of L2 speech act production?
2 Do the type of social situations and L2 proficiency have
differential effects
on the planning speed of L2 speech act production?
3 Do the type of social situations and L2 proficiency have
differential effects
on the speech rate of L2 speech act production?
METHODOLOGY
Participants
The participants were 59 Japanese learners of English in a
Japanese
university. They formed two proficiency groups: 29 higher
proficiency
students (15 males and 14 females, a mean age of 20.48, ranging
from 17
to 25) and 30 lower proficiency students (15 males and 15
females, a mean
age of 19.19, ranging from 18 to 27), based on the institutional
TOEFL scores
(ITP TOEFL) and teacher ratings of oral proficiency. The 8-point
rating scale
118 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
of oral proficiency adapted from ACTFL (1986) and the Ontario
Test of Oral
Interaction Assessment Bands (John 1992) was used to obtain
information
about the learners general speaking ability from classroom
instructors. The
higher L2 group (n 29) had an average ITP TOEFL score of 508,
ranging from480 to 590. Their average oral proficiency rating was
5.48. The lower L2 group
(n 30) had an average ITP TOEFL score of 397, ranging from 330
to 457. Theiraverage oral proficiency rating was 2.56. Except for
one higher learner, none of
the learners had more than six months experience of living
overseas.
Pragmatic speaking task
The pragmatic speaking task measured the learners ability to
understand
situational information and to perform two speech acts, requests
and refusals,
in role plays (see Taguchi 2003, for the details of the task).
This study used
closed role plays, namely a role play in which the actor
responds to the
description of a situation to an interlocutors standardized
initiation (Kasper
2000). Three contextual factors served to categorize two social
situations:
interlocutors power difference (P), social distance (D), and the
degree of
imposition (R) (Brown and Levinson 1987). In one situation type,
the power
relationship was equal, the distance between the interlocutors
was small, and
the degree of imposition was low (PDR-low). In the other
situation type,
the listener had greater power, the interlocutor distance was
large, and the
degree of imposition was also high (PDR-high).
To confirm the PDR-low and PDR-high distinctions, a survey was
carried
out. This survey contained 12 situations and the subjects were
asked
to indicate on a Likert scale from 1 (easy) to 7 (difficult)
according to the
degree of perceived ease/difficulty in performing the target
speech act socially.
The situations were constructed under the influence of five
references: Beebe
et al. (1990), Hudson et al. (1995), Rose and Ono (1995), and
Sasaki (1998).
Based on the existing literature, asking for a pen and refusing
an offer
of coffee were selected as benchmark PDR-low situations. The
rating of 4 on
the scale indicated essentially the same degree of
ease/difficulty with
the benchmark situation. The greater the number was, the more
difficult the
speech act was perceived to be to perform socially, because it
is harder to avoid
giving offense in such a situation. The survey was given to 13
native speakers
of Japanese and 11 native speakers of English in their L1s.1
PDR-high
situations received higher ratings (mean 39.56, SD 2.84) than
PDR-lowsituations (mean 25.04, SD 2.87), indicating that PDR-high
speech actswere perceived to be more difficult to perform socially.
The difference was
significant using a matched t-test (p5 .05). In addition to the
two PDR-lowsituations used as benchmarks, the two situations with
the lowest ratings
were selected for the present task. Among the six PDR-high
situations, four
situations with the highest ratings were selected for the task.
Table 1 displays
the situations used in this study and their ratings.
NAOKO TAGUCHI 119
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
The situations were presented in written form on a card in L1.
Following
Hudson et al. (1995), all role play scenarios had the
participants perform
another speech act in addition to the target one to divert their
attention
away from the particular speech act under study. The task
included one
practice situation and four test situations (see Appendix).
Measures
Three types of measures were used to analyze the learners speech
act
production: appropriateness ratings, planning time, and oral
fluency. The first
measure, appropriateness, was assessed using a six-point rating
scale ranging
from zero to five (Table 2). The sum of the ratings of the four
PDR-high
speech acts (range of 020) and the four PDR-low speech acts
(range of 020)
were used for analysis.
Appropriateness was defined as the ability to perform language
functions
appropriately in a social context and subsumed two elements:
sociopragmatics and pragmalinguistics (Kasper 1992; Leech 1983;
Thomas
1983). By addressing these two holistically, the scale
determined to what
extent learners could use expressions at the proper level of
directness in the
given situations. Grammatical and discourse competencies were
incorporated
into the rating on the basis of the degree to which they
interacted with
appropriateness. Several references served as sources when
developing the
scale (Cohen 1994; North 2000).
Six native speakers, all experienced ESL instructors, evaluated
the speech
acts.2 The interrater reliability was 0.90 for the whole
samples. The samples
that had a large discrepancy in evaluation (i.e. those that were
more than
one point offapproximately 2 percent of the total number of
samples)
were discussed in follow-up meetings, and the average score of
the two raters
was assigned as the final score.
Other measures used in this study aimed to reveal the impact of
task types
on learners speed of pragmatic processing. Two temporal measures
were
Table 1: Role-play situations for the pragmatic speaking
task
Category Situations Mean rating SD
PDR-high Ask your teacher to reschedule the exam. 6.13 0.90
(P D R) Ask your boss to give you a day off. 6.17 0.76Refuse
your bosss request to reschedule work. 6.25 0.74
Refuse the teachers advice to take summer classes. 6.00 0.83
PDR-low Ask your friend for a pen.
(PD I) Ask your sister to pass you the remote. 2.79 1.10Refuse
your friends invitation to the movies. 4.46 1.10
Refuse your sisters offer of coffee.
120 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
used: pre-task planning time and speech rate. Planning time
was
operationalized as the time taken to prepare for each role play
prior to the
task. Speech rate was used as one component of oral fluency that
refers to
fluidity or smoothness of language use (Freed 1995). Speech
rate
(the number of words spoken per minute) has been used
extensively
in the previous research of oral fluency (Ejzenberg 2000; Freed
2000; Lennon
1990; Riggenbach 1991; Towell 2002), as well as in some
task-based studies
(e.g. Robinson 1995, 2001). Previous research also found that
speech rate
positively correlated with other measures of fluency, such as
length of speech
without pauses, hesitations, or repeats (Freed et al. 2004;
Segalowitz and
Freed 2004). In this study, word counts excluded false starts
and repetitions.
Data collection procedures
Data collection took place on campus. After completing a brief
survey and
signing the informed consent form, the participants started the
task. First,
task directions were given by a female native English speaker
interlocutor
who interacted with each learner during role plays. The role
play descriptions
were given via individual situation cards. Every time the
interlocutor gave
a situation card to a participant, she said, Here is the
situation card.
The participants were given an unlimited amount of time to
prepare
mentally. When ready, the participants were instructed to say Im
ready.
and then return the situation card to the interlocutor and begin
the role play.
Planning time was measured from the moment when the interlocutor
said
Here is the situation card. until the moment the participant
said Im ready.
All interactions except the practice one were tape-recorded and
transcribed.
Table 2: Simplified appropriateness rating scale for the
pragmatic speakingtask
Ratings Descriptors
5 Excellent Expressions are fully appropriate.
No or almost no grammatical and discourse errors.
4 Good Expressions are mostly appropriate.
Very few grammatical and discourse errors.
3 Fair Expressions are only somewhat appropriate.
Grammatical and discourse errors are noticeable,but they do not
interfere with appropriateness.
2 Poor Due to the interference from grammatical anddiscourse
errors, appropriateness is difficult to determine.
1 Very poor Expressions are very difficult or too little to
understand.There is no evidence that the intended speech acts are
performed.
0 No performance
NAOKO TAGUCHI 121
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
Data analyses
This study examined whether the two social situation types
(PDR-high and
PDR-low) have a differential effect on features of oral output,
and whether
the effects differed according to the learners L2 proficiency. A
series of
repeated-measures ANOVAs were performed on L2 learners
appropriateness
scores, planning time (i.e. the time taken to prepare for the
role plays),
and the average speech rate (i.e. the number of words produced
per minute).
The between-subject factor was proficiency levels (i.e. high and
low
proficiency groups), and the within-subject factor was the
situation type
(i.e. PDR-high and low). Prior to the statistical analyses, data
distributions
were checked for underlying assumptions. Significance tests for
skewness and
kurtosis and the KolmogorovSmirnov test of normality (Tabachnick
and
Fidell 2001) confirmed the normality of all data distributions,
except that
for planning time and speech rate. Thus, a logarithmic
transformation was
to improve the normality of these data. The significance level
was
set at .05. However, because this study used three statistical
comparisons,
the significance level was adjusted to .017 using the Bonferroni
correction
(Brown 1990).
RESULTS
Tables 3, 4, and 5 display descriptive statistics of
appropriateness scores,
planning time, and speech rate.
Effects of situation types on appropriateness scores
The first research question asked whether learners
appropriateness scores
differed between PDR-high and PDR-low situations and between
groups of
different proficiency. The repeated-measures ANOVA revealed
significant
main effects for situation type (F38.37, p .000, eta square .25)
andL2 proficiency (F 85.20, p .000, eta square .43), and a
significantinteraction between task type and proficiency (F 7.43, p
.009, etasquare .06). Thus, PDR-high speech acts were more
difficult to producethan PDR-low ones. The main effect found for
task type must be interpreted
in light of its significant interaction with proficiency. Lower
proficiency
learners had more difficulty in producing PDR-high speech acts
than higher
proficiency learners.
Effects of situation types on planning time
The second research question addressed the difference in
pre-task planning
speed between PDR-high and PDR-low situation types, and between
the two
proficiency groups. The statistical analyses indicated
significant main effect
122 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
for situation type (F 22.76, p .000, eta square .17). However,
proficiencyrevealed no main effect on planning time (F 2.17, p
.143). In addition,there was no significant interaction effect
between situation type and
proficiency (F .04, p .844). Thus, learners were significantly
faster inplanning when producing PDR-low than PDR-high speech acts.
Because
planning time was not affected by proficiency or by the
interaction between
proficiency and task type, it was concluded that L2 proficiency
had no
significant impact on planning time for both types of speech act
situations.
Effects of situation types on speech rate
The last research question examined the effects of situation
type and L2
proficiency on speech rate. There were significant main effects
for situation
type (F 60.44, p .000, eta square .35) and L2 proficiency (F
46.19,p .000, eta square .29) and a significant interaction between
situationtype and proficiency (F 7.10, p .009, eta square .06).
Thus, learnersshowed significantly faster speech rate when
producing PDR-low than PDR-
high speech acts. The main effect found for situation type must
be
interpreted in light of its significant interaction with
proficiency. The lower
proficiency group was slower in producing PDR-high speech acts
than the
higher proficiency group. A post hoc analysis was conducted to
see whether or
not there is a relationship between the two temporal variables,
namely
Table 3: Production appropriateness scores by speech act
situations types
Group Situation Type K Mean SD Min. Max.
Higher L2 (n 29) PDR-High 4 14.55 2.26 10.00 20.00PDR-Low 4
15.97 2.26 8.00 19.00
Lower L2 (n 30) PDR-High 4 9.20 2.31 4.00 13.50PDR-Low 4 13.08
2.47 7.00 17.50
Notes. Knumber of speech acts. The means show the sum of the
speech act ratings.
Table 4: Production Planning Time (in Seconds) by Speech Act
Situation Type
Group Situation Type K Mean SD Min. Max.
Higher L2 (n 29) PDR-High 4 111.59 69.28 38.83 408.68PDR-Low 4
75.66 31.99 37.62 188.54
Lower L2 (n 30) PDR-High 4 128.32 66.26 44.97 372.31PDR-Low 4
85.50 32.38 40.08 193.69
Notes. Knumber of speech acts. The mean shows the average
planning time per role play.
NAOKO TAGUCHI 123
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
planning time and speech rate. Pearson correlation showed no
significant
relationship, r .04, p 0.8 for higher proficiency group, and r
.05, p 0.8for lower proficiency group.
Descriptive analyses of linguistic expressions in speech
acts
As a post hoc analysis, this study examined linguistic
expressions used by
L2 learners to infer why PDR-high speech acts were more
difficult and took
a longer time to perform than PDR-low ones. Request and refusal
expressions
were classified for their directness levels, using coding
frameworks adapted
from Blum-Kulka et al. (1989) (requests), and Beebe et al.
(1990) and Nelson
et al. (2002) (refusals) (Tables 6 and 7). The frequencies of
all the expression
types were tallied and compared for trends between the two L2
groups.
Native speaker data (10 males and 20 males) elicited under the
same task
conditions was provided as baseline data.
Table 8 displays frequency distributions of request
expressions.
In the PDR-high situations, 100 percent of the native speakers
used
mitigated-preparatory expressions, for example requests embedded
in clause
structures (e.g. Im wondering if clause), while L2 learners,
especiallylower L2 learners, tended to overuse direct expressions,
particularly the
imperatives with please. Learners also relied heavily on hinting
expressions
(39.6 percent), as shown in Example 1. Hints were not
conventionalized,
so they required more inferencing on the part of the
interlocutor and
sometimes more extended negotiations and clarifications. The
percentage
of hinting was greater for lower L2 learners (53.3 percent).
Example 1: L2 learner PDR-high request, asking for a day off
L (learner) : Is it OK if I go to camping with my friend
nextSaturday?
I (interlocutor) : Sure, you can go camping.L : Oh, thank you.
(pause)I : That means you wont be working, right?L : Yeah.
Table 5: Production speech rate by speech act situation type
Group Situation Type K Mean SD Min. Max.
Higher L2 (n 29) PDR-High 4 28.36 13.40 6.89 66.27PDR-Low 4
81.95 27.46 27.75 138.72
Lower L2 (n 30) PDR-High 4 27.73 6.72 11.59 40.01PDR-Low 4 63.17
19.93 37.60 125.22
Notes. Knumber of speech acts. The mean shows the average number
of words produced perminute in each role play.
124 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
Learners overuse of hinting expressions may stem from their
limited
linguistic abilities. The learners in this study did not know
how to express
politeness linguistically in PDR-high situations because they
were not
familiar with mitigated-preparatory expressions that contain
complex
embedded clauses. As a result, they depended on simpler
expressions to
request politely. Their implicit ways of conveying intentions
using hints,
whether successful or unsuccessful, could stem from their
strategies
for being polite and less face-threatening in these formal,
high-stakes
request situations.
Table 6: Request coding framework
I. Direct expressions
1. Imperatives e.g. Please lend me a pen.
2. Explicit performatives e.g. Im asking you to lend me a
pen.
3. Implicit performatives e.g. I want to ask you to lend me a
pen.
3. Obligation statements e.g. You should lend me a pen.
4. Want statements e.g. I want you to lend me a pen.
II. Indirect expressions
II.A. Conventional indirect
5. Preparatory questions e.g. Could you lend me a pen?
6. Suggestions e.g. How about lending me a pen?
7. Permissions e.g. May I borrow a pen?
8. Mitigated preparatory e.g. Im wondering if you could lend me
a pen.
9. Mitigated wants e.g. Id appreciate it if you could lend me a
pen.
II.B. Non-conventional indirect
10. Strong hint e.g. My pen just quit. I need a pen.
11. Mild hint e.g. Can you guess what I want?
Table 7: Refusal coding framework
I. Direct expressions
1. No/Negative willingness/ability e.g. I dont want to./I
cant.
II. Indirect expressions
2. Statement of regret e.g. Im sorry.
3. Wish e.g. I wish I could go.
4. Excuse e.g. I have a plan.
5. Statement of alternative e.g. Id rather drink tea.
6. Promise of future acceptance e.g. Ill do it next time.
7. Indefinite reply/hedging e.g. Maybe we can work something
out.
8. Postponement e.g. Ill think about it.
9. Repetition/question e.g. Friday night?
NAOKO TAGUCHI 125
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
Table 9 displays the frequency distributions of refusal
expressions. Native
speakers used notably more hedging expressions and indefinite
replies
(e.g. probably and kind of). Unlike native speakers, both L2
groups did not
use many indefinite replies or much hedging. Instead, they used
more
apologies, which were almost absent in the native speaker data
(Example 2).
It seems that learners who were unfamiliar with hedging or
indefinite
expressions used apologies as politeness strategies to soften
the refusals.
Example 2: Learner PDR-high refusal, refusing the advice to
takesummer classes
I : Your grades are little low, so you should take summer
sessions.
L : Ah, but I have plan, I have some plan for spring
vacation,summer vacation, so I cant take it.
I : OK.L : Im sorry.I : Just think about it.L : OK.
Table 8: Frequencies of request expressions, PDR-high
situations
NS Higher L2 Lower L2
% (n) % (n) % (n)
I. Direct expressions 0 13.8 (8) 26.7 (16)
1. Imperatives 0 7.0 (3) 21.7 (13)
2. Explicit performatives 0 6.9 (3) 1.7 (1)
3. Implicit performatives 0 0 0
4. Obligations 0 0 0
5. Want statements 0 3.0 (2) 3.3 (2)
II. Indirect expressions 100 (40) 86.2 (50) 73.3 (44)
II. A. Conventional indirect 100 (40) 46.6 (27) 20.0 (12)
6. Preparatory 0 10.3 (6) 10.0 (6)
7. Permissions 0 27.6 (16) 10.0 (6)
8. Suggestions 0 0 0
9. Mitigated-preparatory 100 (40) 6.9 (4) 0
10. Mitigated-wants 0 1.7 (1) 0
II. B. Non-conventional indirect 0 39.6 (23) 53.3 (32)
11. Strong hint 0 39.6 (23) 46.7 (28)
12. Mild hint 0 0 6.6 (4)
Notes. The numbers in the parentheses show the raw counts. There
were 20 native speakers
(NS), 29 higher, and 30 lower learners. The number of requests
analyzed was 40 for native
speakers, 58 for higher, and 60 for lower learners.
126 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
DISCUSSION
This study found that situation type had a significant effect on
both
appropriateness scores and production speed in L2 speech act
production.
PDR-high speech acts were more difficult to produce and required
a longer
production time for learners than did PDR-low ones, regardless
of their
proficiency levels. Thus, the social factors involved in a task,
namely power,
distance, and the degree of imposition, seemed to make certain
situation
types more demanding to perform than others.
The greater degree of difficulty involved in PDR-high speech
acts could
stem from the greater level of politeness required in those
situations.
According to Brown and Levinson (1978), people attempt to
maintain two
types of public image: positive face, the desire to be liked and
approved of by
others, and negative face, the desire to be unimpeded by others.
Refusals and
requests are considered potentially face-threatening because
refusals threaten
the hearers positive face, while requests threaten the hearers
negative face.
Thus, the speaker needs to use various politeness strategies to
minimize such
threats. Longer planning time in PDR-high situations may be a
reflection
of the amount of thinking involved in searching for those
face-saving
expressions. Longer planning time could also reflect learners
lack of
knowledge of those expressions. Because learners did not possess
polite
expressions in their L2 repertoire, it is possible that they
spent longer periods
of time planning how to proceed without them.
This interpretation also corresponds with speech production
processes
discussed in the previous literature (Foss and Hakes 1978; Wiese
1985).
Table 9: Frequencies of refusal expressions, PDR-high
situations
NS Higher L2 Lower L2
% (n) % (n) % (n)
I. Direct expressions 42.9 (30) 39.0 (41) 32.5 (37)
II. Indirect expressions 57.1 (40) 61.0 (64) 67.5 (77)
2. Apology 2.8 (2) 19.0 (20) 22.8 (26)
3. Wish 2.8 (2) 0 0
4. Excuse 37.1 (26) 27.6 (29) 33.3 (38)
5. Alternative 4.3 (3) 2.9 (3) 0.8 (1)
6. Promise for future 4.3 (3) 1.9 (2) 3.5 (4)
7. Indefinite reply/hedging 4.3 (3) 7.6 (8) 1.6 (2)
8. Postponement 0 0 0
9. Repetition/question 1.4 (1) 1.9 (2) 5.3 (6)
Notes. The numbers in the parentheses show the raw counts. There
were 20 native speakers
(NS), 29 higher, and 30 lower learners. The number of refusals
analyzed was 40 for native
speakers, 58 for higher, and 60 for lower learners.
NAOKO TAGUCHI 127
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
Linguistic and non-linguistic knowledge schemata interact in
order to plan
what to say and how to say it. When producing speech acts,
non-linguistic
knowledge, such as speaker relationship and the degree of
formality of the
situations, carries weight in determining appropriate
production. It is possible
that the greater the amount of non-linguistic knowledge
required, as in
PDR-high situations, the longer the planning time becomes. In
the case of L2
learners with limited L2 ability, longer planning time for
PDR-high situations
could reflect their pragmatic challenge as well as linguistic
challenge.
Speech rate, another temporal variable examined in this study,
was also
slower for PDR-high speech acts than for PDR-low ones,
indicating the
influence of situation types on one aspect of oral fluency.
Although both
temporal variables were affected significantly by task type, a
post hoc analysis
revealed no significant relationship between planning time and
speech rate
for both L2 groups, suggesting that these two speed measures
represent
different aspects of performance fluency in speech acts. The two
temporal
variables are affected by the task type, but in an independent
manner.
These results lend support to the two types of planning proposed
by Wendel
(1997): off-line planning (i.e. planning prior to a task) and
online planning
(i.e. planning while performing a task).
The distinction between off-line planning and online planning
was also
documented by Yuan and Elliss (2003) study, which examined the
effects of
pre-task and online planning on story narration by L2 English
learners.
The study found that pre-task planning (planning prior to the
task) enhanced
grammatical complexity, fluency, and lexical variety, while
online planning
(planning while performing a task) promoted accuracy. Different
types of
planning seem to direct learners attention to different aspects
of language,
resulting in different production quality. Pre-task planning
directs learners
attention to the conveyance of message that is reflected in
greater fluency
and lexical variety. Online planning, on the other hand,
encourages learners
attention to grammatical accuracy but results in reduced
fluency.
Combined with previous findings, the present findings imply that
online
and off-line planning engages learners with different
attentional resources.
Learners seem to engage in different thought processes during
performance
and focus on different requirements of a task at different
phases of spoken
production. Because speaking tasks are considered more anxiety
provoking
than other skill areas (e.g. Young 1992), it is possible that
affective factors
such as anxiety, tension, and fear, as well as personality
factors such as risk-
taking, may have influenced planning time in the task. On the
other hand,
cognitive factors such as attention allocation, semantic access,
and
phonological coding may have affected speech rate because
learners were
required to interact spontaneously during role plays.
This study revealed that planning time had distinct
characteristics on its
own, independent of speech rate, general proficiency, and
appropriateness of
speech acts. Despite the more extensive planning time and slower
speech
required for PDR-high speech acts, the mean appropriateness
score for
128 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
PDR-high speech acts was significantly lower compared with
PDR-low speech
acts, which suggests that polite expressions, even when located,
were difficult
to employ. Additional analysis was conducted to check whether
there was
a significant relationship between learners appropriateness
scores and their
planning time. A Pearson correlation revealed no significant
relationship
between the two variables, r0.12, p .71, which provides further
supportfor the present results.
These findings contrast with the previous findings that planning
time
positively influences the quality of learner output (e.g.
Mehnert 1998;
Skehan and Foster 1997; Wigglesworth 1997; Yuan and Ellis 2003).
In this
study longer planning time did not result in more appropriate
speech act
production or more fluent articulation of production. It seems
that planning
time did not necessarily affect the output from the tasks
involving pragmatic
characteristics, probably because those tasks required
processing of different
types of resources, including linguistic, contextual, and
sociocultural
resources. As Mehnert claimed, time reduces the cognitive load
to a
manageable level, but there might be variation in how learners
utilize the
time with respect to the nature of the task.
The difficulty in performing PDR-high speech acts was probably
due to the
relatively complex nature of the appropriate expressions, and
the number
of supporting moves required in production. As shown by the
analysis
of linguistic expressions, for PDR-high requests and refusals,
native speakers
used complex, embedded sentence structures, and mitigated them
with
supporting devices, such as hedges and indefinite responses. For
instance,
in PDR-high requests, 100 percent of the native speakers used
mitigatory-
preparatory expressions (e.g. Do you mind if clause), while they
appearedin less than 10 percent for both L2 groups. These findings
suggest that the
learners did not have the linguistic resources appropriate to
the situation.
In order to compensate for the knowledge gap, L2 learners used
expressions
that were ambiguous and thus ineffective (e.g. hinting) and
consequently
failed to convey illocutionary intent. This tendency was
stronger for the
lower L2 group. Learners spent more time in assessing the
contextual
features in PDR-high situations, but in actual production, they
were limited
in their use of appropriate linguistic forms that map the speech
functions
onto the sociocultural context.
Although the present findings imply that the learners lower
fluency in the
PDR-high situations is a reflection of production difficulty, it
is also possible
that the less fluent speech was evidence of L1 pragmatic
transfer.
In Japanese, a hesitant style is often deemed appropriately
polite in
interactions with higher status interlocutors, especially as
mitigating action
when the situation is face threatening. Speakers tend to avoid
direct,
explicit verbalization of thoughts, and listeners are expected
to become
sensitive enough to read the speakers message (e.g. Okazaki
1993).
Therefore, it is possible that the Japanese participants
transferred
their L1 sociolinguistic norm to their role play interactions in
L2 English.
NAOKO TAGUCHI 129
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
Learners poor performance in PDR-high speech acts, then, seem
to
have resulted from multiple factors, including their overall
low
proficiency, lack of linguistic resources, and transfer of L1
communicative
conventions.
In contrast to PDR-high speech acts, learners found PDR-low
speech acts
easier and faster to produce, probably due to the less
face-threatening nature
of the situation and the lower degree of politeness required.
When producing
PDR-low speech acts, it was unlikely that these speech acts
would threaten
the interlocutors face, and thus there was little need to use
overly polite
expressions. As a result, the expressions used to perform these
speech acts
were relatively short, simple, and thus more easily used by
learners. The
illocutionary force of PDR-low requests was realized by using
fixed linguistic
conventions, such as an utterance Could you verb phrase.
Similarly, fixedresponse patterns such as I cant or No thanks
served as appropriate
refusals in PDR-low situations.
In addition to the linguistic simplicity and conventionality of
these
expressions, learners familiarity with the situations could have
contributed
to the relative ease of performing PDR-low speech acts. It is
also possible
that the PDR-low situations used in this study were more
frequently
experienced in L1 and L2 contexts, and thus were more
contextually
accessible. For these recurrent, standardized situations, the
planning process
may become subconscious and automatic because learners have
available
to them ready-made plans and accessible linguistic means that
reduce
the processing load. Thus, it could be a combination of
accessibility of
linguistic resources, familiarity of situations, and situational
factors that
reduces task demands.
The analyses of the types of linguistic expressions also suggest
the potential
value of looking at strategic devices as criteria when
characterizing learner
performance according to task types. Previous studies used
accuracy, fluency,
and complexity as criteria when analyzing various features of L2
output by
task type. However, as shown in this study, in PDR-high
situations, learners
used various verbal strategies to convey the appropriate level
of politeness.
Thus, the types of linguistic strategies in the learners output
reflected
situational differences.
In summary, this study showed that the different types of
pragmatic tasks
(i.e. PDR-high and PDR-low speech acts) created different
demands on
performance. Some pragmatic tasks were more conventionalized,
immediate,
and were also less face-threatening. Thus, participants were
able to perform
them quickly and easily because they required fewer linguistic
and
psychological resources. In contrast, tasks that were less
common and more
face-threatening were more difficult and took longer to
produce.
The differential task demands that stemmed from these
different
sociolinguistic variables were reflected in the features of the
learners
output, from both a linguistic and processing perspective. The
differential
task demands interacted with learners general proficiency.
130 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
Implications for pragmatic teaching
This study provides evidence for the existence of task types in
pragmatic
functions. It shows that sociolinguistic variables could be
useful criteria in
distinguishing among tasks. Thus, when instructors design tasks
as classroom
activities, they can manipulate both the psycholinguistic and
sociolinguistic
dimensions of these tasks, thereby increasing or decreasing the
tasks
demands and developing tasks of different difficulty levels.
This study also
identified those pragmatic tasks that are relatively easy to
perform and that
should be placed at an earlier stage of the instructional
syllabus. PDR-low
speech acts should be introduced before formal speech acts
because they
were found to be easier and faster for L2 learners to
produce.
Another instructional implication relates to the content of
instruction.
The successful planning and production of speech acts depend on
control
of two dimensions: pragmalinguistics and sociopragmatics (Kasper
1992;
Leech 1983; Thomas 1983). This study demonstrates that,
regardless of
proficiency levels, the learners faced greater challenges with
pragmalinguistics.
In PDR-high situations, they took more time planning, but they
could not use
elaborated linguistic expressions. Thus, for learners similar to
those in this
study, an instructional priority should be placed on
pragmalinguistics,
particularly the choice and control of those language forms in
realizing
speech acts. The expressions gleaned from native speaker data
could serve as
baseline speech act behaviors and help to develop pragmatics
materials.
Limitations and implications for future research
One major limitation is that the study could not reveal the
actual processes that
the L2 learners were engaged in while planning. It was not clear
how learners
were using the planning time; while planning they could have
been searching
for words, thinking about how to avoid grammatical errors, or
looking for
strategies to convey politeness. Future research should use
additional measures
such as think-aloud protocols to examine what aspects of
language production
linguistic or pragmatic aspectsmost concerned the learners while
they were
planning. This line of investigation could help us to understand
the nature of
planning time, which has been shown to bear no relationship with
L2
proficiency, oral fluency, and the appropriateness of speech
acts.
This study is also limited because it used only one measure of
oral fluency,
namely speech rate. The study did not include other measures
such as pause
length, which has been used in previous task-related studies. In
addition,
speech rate measured as words per minute is not sensitive to the
nature of
speech that often involves multi-word lexical chunks. Based on
these
limitations, future research should apply a wider range of oral
fluency
measures to confirm the generalizability of the present
findings.
Another worthy area of investigation is the impact of individual
differences
in production speed. The much greater variance observed in the
time data
NAOKO TAGUCHI 131
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
compared to the score data supports the claim that planning and
production
time involves much individual variability. Individuals tend to
differ in the
ways in which they cope with processing demands in real time.
These
individual differences may be a reflection of differences in
peoples planning
and decision-making, as well as differences in personality,
affective, and
cognitive factors. Speed in pragmatic processing is influenced
by a great
number of non-language factors, including ones sociocultural
knowledge
and experience, which could eventually lead to greater
individual variance.
Because several studies examined the relationship between
personality and
speech production (e.g. Dewaele and Furnham 2000), future
pragmatic
research should incorporate individual differences as additional
variables
when addressing production speed.
Finally, future research should investigate a broader range of
pragmatic tasks
to shed further light on the task difficulty of different
speaking tasks.
Functional ability for language use can be extended beyond the
typical speech
act types so widely investigated to include tasks from other
areas. The areas
suggested as universal pragmatic competence by Kasper and Rose
(2002), such
as face-saving ability and expressions of epistemic stance,
could provide a
useful basis for designing tasks with different sociocultural
characteristics.
Furthermore, in the future research, other measures that are
specific to
pragmatic competence and spoken discourse, such as discourse
management
strategies, can be applied in analyzing learner output. Future
research using a
variety of task conditions drawn from pragmatic and
sociocultural categories, as
well as different types of analytical criteria, could advance
our understanding
of what tasks constitute and how they impact learner
performance.
Final version received November 2005
ACKNOWLEDGMENTS
My sincere thanks to Dr Mary McGroarty, Dr Joan Jamieson, Dr
Susan Fitzmaurice, and
Dr Barbara Freed for their valuable comments and guidance for
this research. I would also like
to thank three anonymous reviewers and the Applied Linguistics
editors for their constructive
comments and suggestions. I am also thankful to Dennis Johnson
for proofreading the
manuscript. A special thanks also goes to Yuko Uchima at Akita
International University for her
assistance on data analyses.
APPENDIX: SAMPLE PRAGMATIC SPEAKING TASK SITUATIONCARDS
PDR-low situation
Its 7:00 PM. You are in the school library studying for
tomorrows English
test. A good friend of yours is also studying in the library.
Your pen just quit,
so you want to ask her to lend you a pen. She asks if you want
to take a
break and go to the movies with her. You want to refuse the
invitation
because you have to study more for tomorrows test.
132 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
PDR-high situation
You are talking with your teacher in her office. Your test is
next Friday, but
you have your friends wedding on the same day. You want to ask
her if you
can take the test at some other time. Your teacher suggests that
you should
take summer session because your grades are low. You want to
refuse the
suggestion because you already have plans.
NOTES
1 The results of the pilot study showed
no discernable differences between
the Japanese and American partici-
pants responses, suggesting that the
PDR-low and PDR-high distinctions
had cross-cultural validity.
2 Experienced ESL instructors were
selected because of their extensive
experience in using holistic assessment
guidelines to evaluate L2 performance.
The raters were asked to listen to each
role play and write the rate that they
thought the most appropriate. After
the norming session, a set of 2025
samples were assigned randomly to
each rater and evaluated indepen-
dently at home. Each set of samples
was evaluated by two different raters.
REFERENCES
Austin, J. L. 1962. How to Do Things with Words.
Cambridge: Harvard University Press.
Beebe, L. M., T. Takahashi, and R. Uliss-Weltz.
1990. Pragmatic transfer in ESL refusals in R.
Scarcella, D. Andersen, and S. Krashen (eds):
Developing Communicative Competence in a
Second Language. New York: Newbury House,
pp. 5574.
Blum-Kulka, S., J. House, andG. Kasper. 1989.
Cross-cultural Pragmatics: Requests and apologies.
Norwood, NJ: Ablex.
Brown, J. D. 1990. The use of multiple t tests in
language research, TESOL Quarterly 24: 7703.
Brown, P., and S. Levinson. 1978. Universals in
language usage: Politeness phenomena in E. N.
Goody (ed.): Questions and Politeness: Strategies in
social interaction. Cambridge: Cambridge
University Press, pp. 56289.
Bygate,M. 1999. Quality of language and purpose
of task: Patterns of learners language on twooral
communication tasks, Language Teaching
Research 3: 185214.
Bygate, M., P. Skehan, and M. Swain. 2001.
Introduction in M. Bygate, P. Skehan, and
M. Swain (eds): Researching Pedagogic Tasks:
Second language learning, teaching, and testing.
New York: Longman, pp. 120.
Candlin, C. 1987. Towards task-based language
learning in C. Candlin and D. Murphy (eds):
Language Learning Tasks. London: Prentice Hall,
pp. 522.
Candlin, C. 2001. Afterward: Taking the curricu-
lum to task in M. Bygate, P. Skehan, and
M. Swain (eds): Researching Pedagogic Tasks:
Second language learning, teaching, and testing.
New York: Longman, pp. 22943.
Chang, Y. F. 1999. Discourse topics and inter-
language variation in P. Robinson (ed.):
Representation and Process: Proceedings of the 3rd
Pacific Second Language Research Forum, Vol. 1.
Tokyo: PacSLRF, pp. 23541.
Cohen, A. 1994. Assessing Language Ability in the
Classroom. Rowley, MS: Newbury House.
Cook,H.M. 2001. Why cant learners of Japanese as
a foreign language distinguish polite from impolite
speech styles? in K. R. Rose and G. Kasper
(eds): Pragmatics in Language Teaching. Cambridge:
Cambridge University Press, pp. 80102.
Council of Europe. 2001. Common European
Framework of Reference for Languages: Learning,
teaching, assessment. Cambridge: Cambridge
University Press.
Crookes, G. 1989. Planning and interlanguage
variation, Studies in Second Language Acquisition
11: 18399.
Crookes, G. and S. Gass. 1993. Tasks in a Pedagog-
ical Context: Integrating theory and practice.
Clevedon, Avon: Multilingual Press.
NAOKO TAGUCHI 133
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
de Bot, K. 1992. A bilingual production model:
Levelts speaking model adapted, Applied
Linguistics 13: 124.
Dewaele, J.-M. and A. Furnham. 2000. Person-
ality and speech production: A pilot study
of second language learners, Personality and
Individual Differences 28: 35565.
Ejzenberg, R. 2000. The juggling act of oral
fluency: A psycho-sociolinguistic metaphor
in H. Riggenbach (ed.): Perspectives on
Fluency. Ann Arbor: University of Michigan,
pp. 287314.
Ellis, R. 1987. Interlanguage variability in narrative
discourse: Style shifting and use of the past tense,
Studies in Second Language Acquisition 9: 120.
Ellis, R. 2003. Task-Based Language Learning and
Teaching. Oxford: Oxford University Press.
Foss, D. J. and D. Hakes. 1978. Psycholinguistics:
An introduction to the psychology of language.
Englewood Cliffs, NJ: Prentice-Hall.
Foster, P. and P. Skehan. 1996. The influence
of planning and task type on second language
performance, Studies in Second Language
Acquisition 18: 299323.
Freed, B. 1995. What makes us think that
students who study abroad become fluent? in
B. F. Freed (ed.): Second Language Acquisition in
Study Abroad Context. Amsterdam: Benjamins,
pp. 12348.
Freed, B. 2000. Is fluency, like beauty, in the eyes
(and ears) of the beholder? in H. Riggenbach
(ed.): Perspectives on Fluency. Ann Arbor: Univer-
sity of Michigan Press, pp. 24365.
Freed, B., N. Seqalowitz and D. Dewey. 2004.
Context of learning and second language
fluency in French, Studies in Second language
Acquisition 26: 277303.
Fulcher, G. and R. Marquez-Reiter. 2003. Task
difficulty in speaking tests, Language Testing 20:
32144.
House, J. 1989. Politeness in English
and German: The functions of please and
bitte in S. Blum-Kulka, J. House, and G.
Kasper (eds): Cross Cultural Pragmatics:
Requests and apologies. Norwood, NJ: Ablex,
pp. 96119.
Hudson, T., E. Detmer, and J. D. Brown. 1995.
Developing Prototypic Measures of Cross-cultural
Pragmatics (Technical Report No.7). Honolulu:
University of Hawaii at Manoa, Second
Language Teaching & Curriculum Center.
John, J. 1992. The Ontario test of ESL oral inter-
action test, System 20: 30516.
Juffs, A. 2001. Psycholinguistically oriented
second language research, Annual Review of
Applied Linguistics 21: 20720.
Kasper, G. 1992. Pragmatic transfer, Second
Language Research 8: 20331.
Kasper, G. 2000. Data collection in pragmatics
research in H. Spencer-Catey (ed.), Culturally
Speaking. London: Continuum, pp. 31641.
Kasper, G. 2001. Four perspectives on L2
pragmatic development, Applied Linguistics 22:
50230.
Kasper, G. and K. Rose. 2002. Pragmatic
Development in a Second Language, Oxford, UK:
Blackwell.
Leech, G. 1983. Principles of Pragmatics. Harlow:
Longman.
Lennon, P. 1990. Investigating fluency in EFL:
A quantitative approach, Language Learning 40:
387417.
Levelt, W. J. M. 1989. Speaking: From intention to
articulation. Cambridge, MA: The MIT Press.
Long, M. and G. Crookes. 1992. Three
approaches to task-based syllabus design,
TESOL Quarterly 26: 5598.
Maeshiba, N., G. Kasper, and S. Ross. 1996.
Transfer and proficiency in interlanguage
apologizing in S. M. Gass and J. Neu (eds):
Speech Acts across Cultures. Berlin: Mouton de
Gruyter, pp. 15587.
McNamara, T. 1996. Measuring Second Language
Performance. London: Longman.
Mehnert, U. 1998. The effects of different lengths
of time for planning on second language perfor-
mance, Studies in Second Language Acquisition 20:
83108.
Nelson, G. L., J. Carson, M. A. Batal,
andW.E.Bakary.2002. Cross-cultural pragmatics:
Strategy use in Egyptian Arabic and American
English refusals, Applied Linguistics 23: 16389.
Norris, J. 2002. Interpretations, intended uses
and designs in task-based language assessment,
Language Testing 10: 33746.
Norris, J., J. D. Brown, T. Hudson, and
J. Yoshioka. 1998. Designing Second Language
Performance Assessments. Technical Report No. 18,
Second Language Teaching and Curriculum
Center: University of Hawaii.
North, B. 2000. The Development of a Common
Framework Scale of Language Proficiency.
New York: Peter Lang.
Nunan, D. 1989. Designing Tasks for the
Communicative Classroom. Cambridge: Cambridge
University Press.
134 TASK DIFFICULTY IN ORAL SPEECH ACT PRODUCTION
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from
-
Okazaki, S. 1993. Stating opinions in Japanese:
listener-dependent strategies in J. E. Atlatis
(ed.): Georgetown University Round Table on
Language and Linguistics. Washington, DC:
Georgetown University Press, pp. 6995.
Ortega, L. 1999. Planning and focus on form in L2
oral performance, Studies in Second Language
Acquisition 21: 10948.
Riggenbach, H. 1991. Towards an understanding
of fluency: Amicroanalysis of nonnative speaker
conversations, Discourse Processes 14: 42341.
Robinson, P. 1995. Task complexity and second
language narrative discourse, Language Learning
45: 99140.
Robinson, P. 2001. Task complexity, task
difficulty, and task production: Exploring
interactions in a componential framework,
Applied Linguistics 22: 2757.
Robinson, P., S. Ting, and J. Urwin. 1995.
Investigating second language task complexity,
RELC Journal 25: 6279.
Rose, K. R. 2000. An exploratory cross-sectional
study of interlanguage pragmatic development,
Studies in Second Language Acquisition 22: 2767.
Rose, K. R. andR.Ono. 1995. Eliciting speech act
data in Japanese: The effect of questionnaire
type, Language Learning 45: 191223.
Sasaki, M. 1998. Investigating EFL students
production of speech acts: A comparison of
production questionnaires and role plays,
Journal of Pragmatics 30: 45784.
Searle, J. R. 1969. Speech Acts: An essay in the
philosophy of language. Cambridge: Cambridge
University Press.
Segalowitz, N, and B. Freed. 2004. Context,
contact, and cognition in oral fluency acquisi-
tion, Studies in Second Language 27: 175201.
Skehan, P. 1995. Analyzability, accessibility, and
ability for use inG. Cook andB. Seidlhofer (eds):
Principle and Practice in Applied Linguistics. Oxford:
Oxford University Press, pp. 91106.
Skehan, P. 1996. A framework for the implemen-
tation of task-based instruction, Applied
Linguistics 17: 3862.
Skehan, P. 1998. A Cognitive Approach to Language
Learning. Oxford: Oxford University Press.
Skehan, P.2001. Tasks and language performance
assessment in M. Bygate, P. Skehan, and
M. Swain (eds): Researching Pedagogic Tasks:
Second language learning, teaching, and testing.
New York: Longman, pp. 16785.
Skehan, P. and P. Foster. 1997. The influence
of planning and post-task activities on accuracy
and complexity in task based learning, Language
Teaching Research 1: 185211.
Skehan, P. and P. Foster. 1999. The influence
of task structure and processing conditions
on narrative retellings, Language Learning 49:
93120.
Swain, M. and S. Lapkin. 2001. Focus on form
through collaborative dialogue: Exploring task
effects in M. Bygate, P. Skehan, and M. Swain
(eds): Researching Pedagogic Tasks: Second language
learning, teaching, and testing. New York:
Longman, pp. 99118.
Tabachnick, B. G. and L. S. Fidell. 2001. Using
Multivariate Statistics (4th edn). Boston: Allyn
and Bacon.
Thomas, J. 1983. Cross-cultural pragmatic
failure, Applied Linguistics 4: 91109.
Thomas, J. 1995. Meaning in Interaction: An
introduction to pragmatics. London: Longman.
Towell, R. 2002. Relative degrees of fluency:
A comparative case study of advanced learners
of French, International Review of Applied
Linguistics 40: 11750.
Trosborg, A. 1995. Interlanguage Pragmatics:
Requests, complaints, and apologies. New York:
Mouton de Gruyter.
Wendel, J. 1997. Planning and second language
production. Unpublished Ed.D. dissertation.
Temple University, Japan.
Wiese, R. 1984. Language production in foreign
and native languages: Same or different? in
H. W. Dechert, D. Mohie, and M. Raupach
(eds): Second Language Productions. Germany:
Gunter Narr Verlag Tugingen, pp. 1125.
Wigglesworth, G. 1997. An investigation
of planning time and proficiency level
on oral test discourse, Language Testing 14:
85106.
Wigglesworth, G. 2001. Influences on perfor-
mance in task-based oral assessments in
M. Bygate, P. Skehan, and M. Swain (eds):
Researching Pedagogic Tasks: Second language
learning, teaching, and testing. New York:
Longman, pp. 186209.
Young, D. J. 1992. Language anxiety from the
foreign language specialists perspective:
Interviews with Krashen, Omaggio Hadley,
Terrell, and Radin, Foreign Language Annals
25:15772.
Yuan, Y. and R. Ellis. 2003. The effects of
pre-task planning and on-line planning on
fluency, complexity, and accuracy in L2
monologic oral production, Applied Linguistics
24: 127.
NAOKO TAGUCHI 135
at University of H
uddersfield on April 12, 2015
http://applij.oxfordjournals.org/D
ownloaded from