The Effects of Task Type
on Monologic Linguistic Features
Shwu-mei Hwang
National Pingtung Institute of Commerce smhuang@npic.edu.tw
Abstract
The use of tasks in language acquisition has been widely discussed and highly promoted in language learning. The question of which tasks influence language development deserves attention. This study was designed to explore how task type influenced particular areas of language use when students performed monologic tasks. 30 junior English majors from one university in southern Taiwan participated in this study, in which they carried out three different oral tasks─narrating, problem-solving, and graph describing. Results showed there were significant differences across different tasks on three analytic measures─ fluency, complexity, and accuracy. As a whole, the participants performed more fluently in the activity of graph description than those of narrative and problem-solving; more syntactic complexity was found in the narrative task; and more accurate speaking quality was revealed in the problem-solving task than the other two. The results shed light on decisions regarding what and when to teach in speaking class. The understanding of the nature of different task demands enables teachers to take an active stance in predicting language patterns before selecting tasks, with a balance in the development of fluency, complexity, and accuracy.
INTRODUCTION
Tasks hold a central place in SLA research and in language pedagogy; they are assumed to engage learners in natural acquisition and help the progress of an interlanguage system toward target forms (Ellis, 1994; Swain, 1995). Different from traditional forms of language teaching, which tend to concentrate on mastery of isolated language structures, the focus of using tasks is on how language development can be promoted when learners are given tasks to perform. As Long and Crooks (1992) pointed out, the importance of a task in instruction is that it allows acquisition process to operate, meaning to be negotiated, and a focus on form to be maintained. Hence, tasks can contribute to language development by setting challenges that require initiation of language in response (Elder, Iwashita, & McNamara, 2002). This “pushed output” plays a diagnostic role in raising learners’ awareness of their strengths and weaknesses, as well as encourages the incorporation of language into learners’ active processing (Bygate, 1999). Skehan and Foster (1997) also pointed out that the “malleable interlanguage system” in learners was modified as a result of participation in tasks.
Previous studies have mostly explored task conditions, especially how planning affected the way learners processed the target language (Foster & Skehan, 1996; Mehnert, 1998; Ortega, 1999; Robinson, 1995; Skehan & Foster, 1997). In addition to language
tasks are to attain a position of importance, it is necessary to know more about the ways tasks themselves influence the nature of performance. Prior literature has also favored conversational interaction, assuming it could best promote language acquisition. Nevertheless, Yuan and Ellis (2003) pointed out that monologic tasks provided a basis for understanding learner performance that is not influenced by interactional, dialogic variables. Moreover, monologues are compatible with most standardized tests in practice. Most importantly, as Brown and Yule (1983) pointed out, tasks that led learners toward monologues could elicit speakers to work to maintain extended turns. Being able to hold a discourse with a greater utterance length is a sign of language progress and is often a goal for EFL speakers.
Language expressions involve creation and choice which may not be completely predictable in advance. Nevertheless, in situations of communication in different genres, tasks can influence the occurrence of linguistic and discourse features (Bygate, 1999). At a view of language focus, different tasks can encourage distinct discourse patterns and certain tasks may be more suitable for achieving particular pedagogical aims. As a result, decisions about what and when to teach certain content become crucial to teachers. Tasks should not be so easy that no extended interlanguage is achieved; tasks should not be so difficult that learners are forced to rely on their lexicalized communication, which may result in early fossilization. Skehan (1996) argued that under cognitive load, three areas─fluency, accuracy, and complexity─get into competition with one another for attentional resources. Balance needs to be established among the three aspects so that development of one should not be at
the expense of another. Too much prioritizing on fluency might lead to overlexicalized performance; consistent concern with accuracy might lead to a lack in fluency and avoidance of using sophisticated language; and excessive focus on complexity might lead to a failure in accuracy (Skehan, 1998). Pedagogical decisions regarding sequencing of tasks can be made by using tasks to target learners’ language structures. Hence, understanding the nature of different task demands can help teachers avoid relying speculation or intuition and further enable teachers to predict discourse patterns. As a result, appropriate tasks can be arranged to direct particular aspects of language and to sequence tasks in order to lead learners to move toward the use of target linguistic features.
LITERATURE REVIEW
Language Processing and Production
In terms of language processing, Levelt (1989) in his book proposed a model concerning three interactive stages ─ conceptualization, formulation, and articulation. The first two phases explain how speakers receive and prepare for speech. In the conceptualization stage, speakers encode the message into propositions, and based on the information, in the formulation stage, speakers construct grammatical and phonological coding of the
they are not able to pay attention to all aspects when receiving and delivering language. This is especially true for EFL learners, whose language competence is generally limited and who need to prioritize what to do, resulting in tension between form and meaning. Under time pressure, if speakers do not need to devote much time to encoding content, they are more likely to attend to form in addition to meaning. On the other hand, if speakers need to sort out how to express their ideas, little attention is left to attend to form. Likewise, if tasks contain more “dense” information (e.g., nominal information, specific terms), speakers need to spend more time encoding and organizing the propositional content. They will pay less attention to form than the case in performing tasks with less density.
Task Types in Relation to Task Difficulties
Studies have revealed that learners’ discourse patterns are influenced by task structures (Shohamy, 1994; Young, 1995; Young & Milanovic, 1992). Brown, Anderson, Shilcock, and Yule (1984) in their book proposed a matrix, pointing out the level of task difficulty ascends from static (e.g., a diagram) to dynamic (e.g., telling a story) to abstract (e.g., expressing opinions) task. In other words, with fixed, visual information, it is relatively easy for speakers to get the job done. However, when more sequencing events and causal relationships are involved in conveying a message, the task becomes more demanding. Furthermore, tasks involving non-existent, self-manipulating elements are the most difficult. Another dimension, they pointed out, in influencing task difficulty is related to scale and interrelationships of elements. Not surprisingly, more elements or characters lead to greater task difficulty. Robinson, Ting, and Urwin
(1996) proposed that task difficulty is influenced by the dimensions of cognitive load, planning time, and prior information, concluding that a task involving a “there and then” narrative without an access to immediate visual support represents a greater processing burden and is harder than a “here and now” task where speakers have such access and support. A study by Foster and Skehan (1996) showed that tasks based on personal, more familiar information were easier than those with unfamiliar and more remote information. In sum, tasks that involve pictures are fixed in nature and have familiar information tend to be easier.
Tasks in Relation to Fluency, Accuracy, and Complexity
Loschky and Bley-Vroman (1993) showed that different tasks required speakers to employ linguistic features at different degrees. More demanding tasks required more attention to task transaction, which resulted in limited attention available for focus on form. On the contrary, if speakers have well-organized chunks of knowledge ready for task performance, they may have additional control over other language demands. Tong-Fredericks’ (1984) study found that open tasks generated more accuracy and complexity because no single correct solution was required. Similarly, Brown (1991) found that an interpretative task which was open in nature resulted in more complex language use than a closed decision-making task. Robinson (1995) found that oral narratives were more fluent in a “here and now”
to ensure all the necessary information were included in the message. Similarly, Iwashita, McNamara, and Elder (2001) found that a “there and then” condition led speakers to produce more accurate language than “here and now.” Studies by Foster and Skehan (1996) and Skehan and Foster (1997) on three tasks ─ personal information exchange, narrative, and decision-making tasks─found that personal information exchange and decision-making tasks led to significantly higher accuracy than the narrative ones, while the personal information exchange task led to lower complexity than the other two tasks. The narrative and decision-making tasks generated significantly less fluency than the personal tasks. Language output was the most complex and least accurate in narrative tasks. Skehan and Foster (1999) also explored the effects of structure and processing load on a narrative retelling task and found that more structured tasks generated more fluent language, while a heavier processing load led to complexity of language. To summarize, fluency is likely to be reached in more structured tasks or when speakers obtain visual support; complexity is linked to heavier processing load or when there is no single correct solution; accuracy is expected when familiar information is involved or when speakers do not obtain visual input support. For the limited processing capacity of speakers, it seems that the gain in one aspect of language production may be at the loss of another.
This current study aims to explore how different tasks containing different levels of perceptual and cognitive demands affect speakers’ linguistic features in terms of fluency, accuracy, and complexity. Overall, this paper aims to explore the value of prediction on the use of different tasks so that classroom learning can be directed
by the use of tasks. Moreover, the understanding of how learners’ language use is influenced when performing a particular type of task can shed light on directing learners to comply with linguistic requirements of fluency, accuracy, and complexity.
The research questions of this study are as follows: 1. To what extent does speakers’ oral production across
tasks differ in fluency?
2. To what extent does speakers’ oral production across tasks differ in complexity?
3. To what extent does speakers’ oral production across tasks differ in accuracy?
METHOD
Design
This is a single factor within-subjects design. Each participant performed three different types of tasks (narrative, problem-solving, and graph description) so that the researcher could examine how task features influence language production. The three types were monologic in nature in this current study. In order to counterbalance the practice effect, participants were randomly divided into three groups and each group completed the three tasks in different order. The order for presenting tasks is listed in Table 1.
Table 1
Task Order Across the Groups
Groups Tasks
Group 1 narrative graph description problem-solving Group 2 problem-solving narrative graph description Group 3 graph description problem-solving narrative
Participants
Participants in this study consisted of 30 juniors majoring in English at a university in southern Taiwan. They had had formal English learning since junior high school and were quite a homogeneous group in terms of educational background and age (around 21 to 23). They had taken a simulation TOEFL-ITP (Test of English as a Foreign Language-Institutional Testing Program) test before the study. Their scores ranged from 450 to 530, a proficiency level that could be interpreted as intermediate-low to intermediate.
Materials
The three tasks used were narrative, problem-solving, and graph description. These choices were based on an analysis of tasks commonly used in Business English textbooks. Passages similar to topics on narration, problem-solving, and graph description were chosen and altered. In order to reach face validity, the first draft of the task materials was examined by two English teachers. Upon receiving their comments, the researcher made some modifications and the draft for the current study was devised. Narration was a business event
regarding piracy and intellectual property rights in China containing 121 words; problem-solving was the impact of globalization on a factory decrease due to its poor quality control and lack of creativity in design, containing 93 words in length; and the graph description was the stock exchange volumes in three stock markets in four years (see Appendix). Figure 1 is the presentation of the three tasks.
Task Features
Of the three tasks, the narrative task requires participants to describe an event. Such a task involves encoding new information into linguistic form; memory load is involved and greater cognitive effort is required to carry out this type of task. Therefore, greater complexity in language production is expected. The problem-solving task involves addressing a problem with no definite answer. Such an “open” type of task is not as cognitively demanding as the narrative one. More complex and accurate language production can be anticipated. With participants’ access to the graphical, visual input, the static characteristic suggested that the graph description should be the least cognitively demanding and the easiest. The most fluent language would occur in this task. Distinct characteristics involved in the three types of tasks are described in Table 2.
Figure 1
Table 2
Distinct Characteristics in the Three Tasks
Task Type Description Interpretation Narrative Participants are asked to
talk about the story after they read it.
-the most cognitively demanding
-participants have to first comprehend and then organize detail of the story Problem-solving Participants are asked to
offer solutions to the problem a company faces.
-less cognitively demanding than narratives but more demanding than graphs
- participants have to give reasons to justify decisions they have made Graph description Participants are asked to
interpret the data.
-the least cognitively demanding
-participants are assisted with pictorial display of information and clear structure
Measures
The measurement of oral production has long been problematic in that different measures are adopted in different studies (Ellis, 2005). The criteria applied in this study came from similar studies concerning fluency, complexity, and accuracy.
Fluency. If more utterances occur in a limited time, we assume more fluency is exhibited. Some non-fluency features such as hesitations, pauses, false starts, reformulation, and repetitions are common in oral production. By obtaining information regarding total and meaning syllables within a fixed time (speech rate), we can
utterances with no meanings such as “m~m” and unfinished utterances such as “prob”.
Complexity. If more syntactic varieties such as tense (e.g., present, past, present perfect), modality (e.g., should, can, must), and voice (e.g., active, passive) are used in the production, the task load are heavier. The T-unit (a main clause plus any other clauses which are dependent upon it; Foster, Tonkyn, & Wigglesworth, 2000) was used as it was a more suitable measure for this study since monologic tasks contain fewer omitted utterances (Foster et al., 2000; Yuan & Ellis, 2003). For example, “The story is about how the EU trade chief urges China to open its market and control piracy.” is considered one T-unit. Complexity measures comprised the number of different grammatical forms used in the task, including tense, modality, and voice (Foster & Skehan, 1996; Yuan & Ellis, 2003). For example, “They can also increase advertising to promote their products and that will improve their reputation.” contains one modality and two different tense types (present and future).
Accuracy. Accuracy was measured according to syntactic correctness. Accuracy measures contained, first, error-free clauses─ the percentage of clauses that do not contain any error─and second, correct verb forms─the percentage of correct verbs in terms of tense, modality, and subject-verb agreement (Foster & Skehan, 1996; Yuan & Ellis, 2003).
Procedures
This study, which was semi-direct speaking assessment in nature, was conducted in October 2007. The students performed the three tasks in a language laboratory during a Business English class.
Instructions were given to students regarding how to record their speaking, and the participants had trials before their actual performance to minimize influence from operation-related factors. Stimulus materials were presented to students via computers. Participants were given 10 minutes to prepare once the texts were shown on the computer screen. The choice of planning time was based on previous studies to ensure measurable effects on language in fluency, accuracy, and complexity (Foster & Skehan, 1996; Mehnert, 1998; Yuan & Ellis, 2003). In their preparation period, students were allowed to choose whatever strategy they thought was helpful to them (such as taking notes or rehearsing orally) and to practice individually. Moreover, they were informed that notes would be collected before their recording. They were required to spend no more than 3 minutes on each task. The audiotaped data were transcribed and coded. Ten pieces of oral output along with the transcripts from each task were randomly selected and were checked by two native English teachers to ensure that the two matched.
Data Analysis
The independent variable in this study was task type, and the dependable variables were fluency, complexity, and accuracy measures in speaking performance. Repeated measures ANOVA was performed to examine whether there were linguistic differences across the three tasks. Moreover, post-hoc analyses were carried out to
RESULTS
Fluency
With regard to fluency measured by total syllables in one minute, results showed that of the three tasks, participants uttered the most syllables when performing the graph description task and the fewest syllables in the narrative task. Means in the narrative, problem-solving, and graph description tasks were 96.88, 107.67, and 132.89, respectively. With regard to fluency measured by meaningful syllables per minute, the results still showed that participants uttered the most syllables when performing the graph description task and the fewest syllables when performing the narrative task. Means in the narrative, problem-solving, and graph description tasks were 83.61, 98.04, and 132.89, respectively. Descriptive statistics of fluency in the participants’ performance when carrying out the three tasks are shown in Table 3.
Table 3
Descriptive Statistics of Fluency in the Three Tasks
Task Type (N = 30)
Fluency (per minute)
Total Syllables Meaningful Syllables
M SD M SD
Narrative 96.88 22.37 83.61 23.02 Problem-solving 107.67 22.57 98.04 24.23 Graph Description 132.89 27.33 124.82 28.79
When repeated measures ANOVA and post-hoc analyses were performed, results showed that participants had 36 more syllables in the graph description task than in the narrative task (p < .000); and 25 more syllables in the graph description task than in the problem-solving task (p < .000). Moreover, results showed that participants produced 14 more meaningful syllables in the problem-solving task than in the narrative task (p = .014); 41 more meaningful syllables in the graph description task than in the narrative task (p < .000); and 26 more meaningful syllables in the graph description task than in the problem-solving task (p < .000). This indicates that participants spoke faster in the graph description task than the other two tasks, no matter whether total or meaningful syllables were concerned, and thus demonstrated the most fluent discourse. Differences in fluency across the three tasks are shown in Table 4.
Table 4
Differences in Fluency Across Tasks
Task Comparisons MD SE Sig.
Total syllables
Narrative vs. Problem-solving -10.79 5.82 .074 Narrative vs. Graph Description -36.01 6.02 .000* Problem-solving vs. Graph Description -25.22 4.89 .000*
While there was no significant difference in total syllables produced between the narrative and the problem-solving task, there were significant differences regarding meaningful syllables between the two, indicating there were more hesitations and repetitions when students performed the narrative task. Moreover, when participants’ total and meaningful syllables were compared, paired-samples t-tests showed there were significant differences between the two (p < .000). Students had more hesitation, reformulation, and repetition in the total syllables produced across the three tasks. Of the three tasks, participants had the most syllable differences in the narrative task and the fewest syllable differences in the graph description task. The t values in the narrative, problem-solving, and graph description tasks were 9.37, 7.26, and 5.24, respectively. Differences in the total and meaningful syllables produced across the three tasks are shown in Table 5.
Table 5
Differences in Total and Meaningful Syllables Produced Across the Three Tasks
Task Type (N = 30) Differences in Syllables Produced
M SD t df Sig. Narrative Total syllables 96.88 22.37 9.37 29 .000* Meaningful syllables 83.61 23.02 Problem-solving Total syllables 107.67 22.57 7.26 29 .000* Meaningful syllables 98.04 24.23 Graph Description Total syllables 132.89 27.33 5.24 29 .000* Meaningful syllables 124.82 28.79 Note. *p < .05
Complexity
The syntactic complexity was measured in terms of tense, modality, and voice. Results showed that of the three tasks, participants produced the most tense variety when performing the narrative task, the most modality variety when performing the problem-solving task, and the most voice variety when performing the graph description task. Descriptive statistics of complexity in participants’ performance when performing the three tasks are shown in Table 6.
Table 6
Syntactic Complexity in Task Completion
Complexity (T-unit Clause)
Task Type (N = 30) Tense Modality Voice
M SD M SD M SD
Narrative 2.50 .90 .50 .51 .43 .57
Problem-solving 1.60 .67 1.77 .82 .10 .31 Graph Description 2.03 .89 .17 .38 .70 .65
When repeated measures ANOVA and post-hoc analyses were performed, results indicated there were significantly more tense use in the narrative task than in the problem-solving task (p < .000) and the graph description task (p = .024). With regard to modality, there was
task than in the graph description task (p = .005). With regard to voice, there was significantly more voice variety in the narrative task than in the problem-solving task (p = .005). Moreover, there was significantly more voice variety in the graph description task than in the problem-solving task (p < .000). Differences in complexity across the three tasks are shown in Table 7.
Table 7
Differences in Complexity Across Tasks
Task comparisons MD SE Sig.
Tense
Narrative vs. Problem-solving .90 .09 .000* Narrative vs. Graph Description .47 .20 .024* Problem-solving vs. Graph Description -.43 .23 .068
Modality
Narrative vs. Problem-solving -1.27 .20 .000* Narrative vs. Graph Description .33 .11 .005* Problem-solving vs. Graph Description 1.60 .14 .000*
Voice
Narrative vs. Problem-solving .33 .11 .005* Narrative vs. Graph Description -.27 .17 .133 Problem-solving vs. Graph Description -.60 .14 .000*
Note. *p < .05
Accuracy
Accuracy was measured in two approaches─error-free clauses and correct verb forms. Results showed participants demonstrated a
higher percentage of correctness in verb forms than error-free clauses. Of the three tasks, participants displayed the most correct linguistic performance in the problem-solving task (M = .48), followed by the graph description task (M = .47), and then the narrative task (M = .31). Participants produced the most correct verb forms in the problem-solving task (M = .83), followed by the graph description task (M = .64), and then the narrative task (M = .56). Descriptive statistics of the accuracy in participants’ performance when performing the three tasks are shown in Table 8.
Table 8
Statistical Description of Accuracy in the Three Tasks
Task Type (N = 30)
Accuracy (Clauses produced) Error-free Clauses Correct Verb Forms
M SD M SD
Narrative .31 .25 .56 .26
Problem-solving .48 .23 .83 .17
Graph description .47 .19 .64 .20
When repeated measures ANOVA and post-hoc tests were performed, results indicated participants had significantly more error-free clauses in the problem-solving task than in the narrative task (p = .001). Moreover, there was significantly more correct
in the narrative task (p < .000), and the graph description task (p = .001). Differences in accuracy across tasks are shown in Table 9.
Table 9
Differences in Accuracy Across Tasks
Task comparisons MD SE Sig.
Error-free Clauses
Narrative vs. Problem-solving -.17 .04 .001* Narrative vs. Graph description -.16 .05 .007* Problem-solving vs. Graph Description -.01 .06 .855
Correct Verb Forms
Narrative vs. Problem-solving -.27 .05 .000* Narrative vs. Graph description -.08 .06 .243 Problem-solving vs. Graph description .19 .05 .001*
Note. *p < .05
Summary of Major Findings
Results revealed different levels of task effects on participants’ oral production in the three linguistic measures. Regarding the effects of task type on the oral production of participants in each of the tasks, the major findings are summarized as follows:
1. As far as fluency is concerned, as a whole, participants had more fluent utterances in the graph description than the narrative and problem-solving tasks.
2. Regarding complexity, participants employed different combinations of syntactic variety in completing the three tasks. As a whole, they used more tense and voice variety in
the narrative task.
3. As for accuracy, as a whole, participants produced more accurate clauses and verb usages in the problem-solving task than in the narrative and graph description tasks. Table 10 summarizes the significant differences in participants’ language production in the three tasks.
Table 10
Summary of Significant Differences in Oral Production in the Three Tasks
Tasks Fluency Complexity Accuracy
TS MS T V M E-F C CVF Narrative N>P N>G N>P N>G Problem-solving P >N P >N P >G P >N P >N P >G Graph Description G>N G>P G>N G>P G>P G>N
Note. TS = total syllables; MS = meaningful syllables; T = tense; V = voice; M = modality; E-F C = error-free clauses; CVF = correct verb forms; N=
DISCUSSION
Fluency
In this current study, significant differences were found in task types on all three analytic measures ─ fluency, complexity, and accuracy. Regarding task difficulty, it was predicted that the graph description task would be the easiest, while the narrative task would be the most cognitively demanding. The results confirmed this prediction. In terms of fluency, the results indicated that the graph description indeed elicited much more fluent performance, as indexed by the number of syllables per minute. This suggested that being able to draw upon familiar, ready-encoded information does promote a greater degree of fluency regarding the total and meaningful syllables produced in completing tasks. Of the three tasks, the task with visual information was shown to facilitate language processing and production. The illustration task greatly assisted test-takers’ oral fluency, as its content and differences were immediately revealed to language users without much processing load. The results indicated that when the major points were packed in relatively fewer visual chunks, more fluent language production was produced. The most fluent language production was associated with a low level in complexity and moderate level in accuracy. Moreover, the significant differences between total and meaningful syllables produced among participants indicated that even though speakers were more fluent in performing the graph description task, significant redundancies and pauses were revealed in their discourse, thus suggesting the need for more training in this aspect.
Complexity
In terms of complexity, the results indicated differences in the combination of syntactic variety. While participants showed more overall language complexity in the narrative task, they had more modality use in the problem-solving task than in the narrative and graph description tasks. Moreover, students had more use of voice in the graph description than in the problem-solving task. It indicated that first, participants were quite aware and were able to present syntactic variation according to the demand of task types (e.g., more modality use in the problem-solving task and more tense variety in the narrative task). Second, the narrative task was found to have greater processing load since speakers were asked to devote much attention to encode the propositional content from the written text, absorb the main points, and express what they had read. In order to meet the demands of this task, learners were forced to make more effort in this regard, with little capacity available to attend to other aspects. The narrative task required students to use more tense and voice variety in their expression, which resulted in more language complexity. This created conditions for discourse expansion and development in complexity at the expense of the lowest linguistic fluency and accuracy.
Accuracy
participants to produce discourse without errors. In fact, the number of error-free clauses was very limited, with the means of percentage all falling under .50, indicating students had least attention devoted to accuracy and their production of correct clauses needed to be enhanced. Overall, participants performed better in the problem-solving task than in the narrative and graph description tasks. Since more use of modal verbs (can, should) could be expected in a problem-solving task, speakers did not have to pay attention to the grammatical changes in verbs, and this could be a reason their levels of accuracy increased. Another reason might be that the use of fixed chunks of phrases with explicit sentence structures was more frequent in the problem-solving task, and thus contributed to accuracy. Overall, the most accurate language production was associated with a moderate level in fluency and a low level in complexity.
The nature of some tasks may be more efficient in eliciting certain targeted language features. The graph description generated a greater level of accuracy without achieving much complexity. The narrative task produced the highest level of complexity, with a greater need to use precise and extending language. Greater complexity in linguistic performance seemed to be at the expense of accuracy and fluency. The problem-solving task elicited greater accuracy with a moderate level in fluency and a low level in complexity. It was apparent that with different task types, under information-processing pressure, participants allocated attention to particular goals. Since the narrative task asked participants to remember more factual information, it was more difficult for speakers to maintain the speech flow over long periods. This was why students often stopped in their speech as they tried to come up with ideas for expression; more
hesitation and reformulations were found, which resulted in low levels of fluency. The problem-solving task pushed learners to develop a certain brainstorming quality, and students’ sentences tended to be accurate but short. While participants were able to speak fluently in the graph description task, their language tended to be clear but simple, containing fewer variations in discourse patterns.
CONCLUSION
Apparent distinctions were revealed among the three tasks─ narrative, problem-solving, and graph description ─ with linguistic differences generated in complexity, accuracy, and fluency, respectively. With a clear decoding message, the graph description task can be employed to encourage fluent output. A problem-solving task which contains a clear structure can be adopted to encourage language accuracy. Moreover, the propositional elements in a narrative task require speakers to utilize syntactic variety for message conveyance. Similar to the previous studies (Foster & Skehan, 1996; Skehan & Foster, 1997; Skehan & Foster, 1999), data in the present study implied that a heavier processing load (narrative) generated greater complexity; a clearer inherent structure (problem-solving, graph description) favored either accuracy or fluency. Since task features can have an effect on the nature of performance, the understanding of how variables
attained. Relevant to pedagogic contexts, tasks can be manipulated to maximize the chances that learners’ attention will be directed. If fluency is a pedagogic goal, then clearly the use of familiar, visual input is indicated. If complexity is important, a heavier processing load would seem to be the best choice. Consequently, pedagogical decisions regarding the sequencing of tasks can be made by using tasks to elicit learners’ language structures appropriate to the specific developmental stage.
Language use and development is a continuous interplay between meaning and form, and there are some implications for the use of task types in instruction based on this study. Tasks can be employed to provide a basis for students to work on their language in which speakers are pushed to make language choices in their expression. Tasks serve to provide opportunities for teachers to achieve particular instructional goals. There was evidence that different tasks elicited different patterns of language use. The three different types of task guided students in different directions and styles in their language quality. For teachers, evaluating and sequencing tasks involves knowledge regarding the degree of difficulty associated with language processing and production, grammar, and vocabulary use, as well as the knowledge of the learner’s background. And understanding task characteristics is essential to pedagogical decisions regarding the assessing and sequencing of tasks in syllabus design. Rather than simply observing results after task completion, teachers can predict task quality and value in advance to make them as sources in instruction.
A number of limitations of this study are evident. First, this was a small-scale study with 30 participants. A larger number of
participants might reveal clearer conclusions regarding linguistic differences in the three measures. Second, the statistical analysis was based on the random selection of participants, which was not fulfilled in this study, although random assignment of participants to the three tasks was carried out. Third, the participants in this study were at a pre-intermediate to intermediate language proficiency level, and they may have different language abilities from learners at a lower or higher proficiency level. Hence, sequences regarding these tasks may vary when participants’ proficiency levels differ. Finally, participants carried out these tasks through the integration of reading and writing. If aural input were included, the language quality may be different.
ACKNOWLEDGEMENTS
I am most grateful to participants in this study, and anonymous reviewers, whose comments and advice have greatly improved this paper.
REFERENCES
Brown, G. (1991). Group work, task difference, and second language acquisition. Applied Linguistics, 21, 1-12.
talk: Strategies for production and assessment. Cambridge, UK: Cambridge University Press
Bygate, M. (1999). Quality of language and purpose of task: Patterns of learners’ language on two oral communication tasks.
Language Teaching Research, 3, 185-214.
Elder, C., Iwashita, N., & McNamara, T. (2002). Estimating the difficulty of oral proficiency tasks: What does the task-taker have to offer? Language Testing, 19, 347-368.
Ellis, R. (1994). The study of second language acquisition. Oxford, UK: Oxford University Press.
Ellis, R. (2005). Task-based language learning and teaching. Oxford, UK: Oxford University Press.
Foster, P., & Skehan, P. (1996). The influence of planning time and task type on second language performance. Studies in Second
Language Acquisition, 18, 299-323.
Foster, P., Tonkyn, A., & Wigglesworth, G. (2000). Measuring spoken language: A unit for all reasons. Applied Linguistics, 21, 354-375.
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, 401-436.
Levelt, W. J. M. (1989). Speaking: From intention to articulation. Cambridge, MA: Massachusetts Institute of Technology Press. Loschky, L., & Bley-Vroman, R. (1993). Grammar and task-based
methodology. In G. Crookes, & S. M. Gass (Eds.), Tasks and
language learning: Integrating, theory and practice (pp. 123-167). Clevedon, UK: Multilingual Matters.
Long, M., & Crooks, G. (1992). Three approaches to task-based syllabus design. TESOL Quarterly, 26, 27-56.
Mehnert, M. (1998). The effects of different lengths of time for planning on second language performance. Studies in Second
Language Acquisition, 20, 83-108.
Ortega, L. (1999). Planning and focus on form in L2 oral performance.
Studies in Second Language Acquisition, 21, 109-148.
Robinson, P. (1995). Task complexity and second language narrative discourse. Language Learning, 45, 99-140.
Robinson, P. S., Ting, C. -c., & Urwin, J. J. (1996). Investigating second language task complexity. RELC Journal, 26, 62-79. Shohamy, E. (1994). The validity of direct versus semi-direct oral
tests. Language Testing, 11, 99-124.
Skehan, P. (1996). A framework for the implementation of task-based instruction. Applied Linguistics, 17, 38-62.
Skehan, P. (1998). A cognitive approach to language learning. Oxford, UK: Oxford University Press.
Skehan, P., & Foster, P. (1997). Task type and task processing conditions as influence on foreign language performance.
Language Teaching Research, 3, 185-211.
Skehan, P., & Foster, P. (1999). The influence of task structure and processing conditions on narrative retellings. Language Learning,
49, 93-120.
Tong-Fredericks, C. (1984). Types of oral communication activities and the language they generate: A comparison. System, 12, 133-145.
Young, R. (1995). Conversational styles in language proficiency interviews. Language Learning, 45, 3-42.
Young, R., & Milanovic, M. (1992). Discourse variation in oral proficiency interviews. Studies in Second Language Acquisition,
14, 403-424.
Yuan, F., & Ellis, R. (2003). The effects of pre-task planning and on-line planning on fluency, complexity and accuracy in L2 monologic oral production. Applied Linguistics, 24, 1-27.
ABOUT THE AUTHOR
Shwu-mei Hwang received her PhD from Illinois State University, Normal, Illinois. She is an assistant professor in the Applied English Department at National Pingtung Institute of Commerce. Her research interests include exploring the potential pedagogical options for improving oral ability in the EFL context and employing tasks to the analysis of oral production.
APPENDIX
Narrative Task
The EU trade chief urged China to open its markets to EU goods and control piracy and counterfeiting. Mr. Mandelson said on Friday after talks with Chinese Trade Minister Bo Xilai that Europe needs to keep the markets open to fair trade from China. It also means that China has to open its markets more to European exports, and of course enforce intellectual property rights.
Earlier this year, trade relations between the EU and China came to disagreement during a fight over the quantity of textiles China exports to Europe. And the tension has become higher as Chinese shoes have been sold to EU at unfairly low prices. In addition, Europe's car industry has complained that Beijing is restricting European activity in China.
Problem-solving Task
Fortune Garments is one of the oldest trading groups in Hong Kong, making high-quality clothes. It has become a global market with sales of over 1.8 billion. Globalization has brought problems in the company’s overseas factories. Among them are poor quality control (order cancellation is frequent and Fortune Garments has lost lots of sales), and lack of creativity in design (its latest design was described as boring and behind the times).
Graph Description Task
The following graph is the total capital of exports ($ million) in three cities. How would you describe it?
0 1 2 3 4 5 2001 2002 2003 2004 New York Tokyo London
