tefl papers

ISSN 1799-2591
Theory and Practice in Language Studies, Vol. 1, No. 11, pp. 1549-1560, November 2011
© 2011 ACADEMY PUBLISHER Manufactured in Finland. doi:10.4304/tpls.1.11.1549-1560 The Effect of Output Tasks on the Acquisition of
English Verbal Morphemes
Shirin Abadikhah
University of Mazandaran, Iran
Email: [e-mail]

Foroogh Zarrabi
University of Mazandaran, Iran
Email: [e-mail]
Abstract—This paper aims to explore the facilitative effect of an output task on learning verbal morphemes by
Iranian EFL learners. The main research question addressed whether engaging learners in output tasks can fill the gap between comprehension and production of verbal morphemes (present progressive–ing and simple past–ed) and promote learning of the target linguistic forms. Two groups of learners from young adult classes
(all male students within the age range of 13-15) participated in the current study: an output group consisting of 20 learners and a control group consisting of 18 learners. All of them were administered a pretest and posttest. The learners in the output group worked on a text reconstruction task which employed an audio-text cartoon strip. The major findings were (a) the comparison of pre-test scores showed a significant difference between comprehension and production of target forms by the learners, (b) after receiving treatment, the output group outperformed the control group, (c) finally, the control group failed to show comparable improvement in their production although they had measurable gains in comprehending the target linguistic form. Considering the above findings, further support was found for the effect of output tasks on L2 learning.
The findings contribute to the research that has examined type of task and level of processing in promoting L2 learning. Index Terms—output tasks, verbal morphemes, reconstruction task

I. INTRODUCTION
Finding the most efficient way to teach grammar has been one of the most controversial issues in SLA over the past few decades (Celce-Murcia, 1991; Ellis, 2006). In search for the best way to teach grammar, the roles of input and output have received substantial attention in second language acquisition (SLA) theory and numerous studies have produced major insights in the field of SLA. Earlier studies gave the idea that acquisition is a natural outcome of comprehensible input (Krashen, 1985). Such studies have chiefly focused on the significance of comprehensible input in developing learners‟ knowledge of the target language. Lately, however, several studies have suggested that the role of output is as important (if not more) for acquisition of a second language as the role of input. These studies‟ claim is mainly based on Swain‟s output hypothesis (Swain, 1985) which considers output as the cause of L2 acquisition, not just the product of it. Studies on output (e.g. Dekeyser & Sokalski, 1996; de la Fuente, 2006; Izumi, 1999; Izumi, 2002;
Song & Suh, 2008) together with formal and informal observations of Canadian immersion program (Swain, 1985) provide empirical evidence that developing productive ability of learners and language acquisition requires more than mere comprehending the language.
Many research findings to date have explored the role of output and different functions of it in language learning
(Bygate, 1999; DeKeyser & Sokalski, 1996; De la Fuente, 2006; Garcia Mayo, 2002; Geeslin, 2006; Izumi, 2002; Izumi
& Bigelow, 2000; Izumi et al., 1999; Izumi & Izumi, 2004; Newton & Kennedy,1996; Kim, 2009; Kuiken & Vedder,
2008; Lyster, 2004; Mehrang & Rahimpour, 2010; Mennim, 2003; Nobuyoshi & Ellis,1993; Qin, 2008; Reinders, 2009;
Shehadeh, 1999; Shehadeh, 2003; Song & Suh, 2008; Storch,1998a &1998b; Swain & Lapkin, 1995; Vickers & Ene,
2006; Yoshimura, 2006). Among the various functions of output proposed by Swain (1995), the noticing function has received substantial attention since many research findings show that noticing and attention play significant roles in language learning. Nevertheless, the studies which specifically have explored whether output tasks promote learning of target linguistic forms, as a result of noticing function, have produced mixed results (Izumi, 2002; Izumi & Bigelow,
2000; Izumi et al., 1999; Izumi & Izumi, 2004; Song & Suh, 2008).Therefore, more research needs to be done to explore the issue. Besides, the possible effect of these tasks on learning grammatical morphemes has not been fully explored. This is despite the fact that many learners have problems in using these morphemes in natural settings – particularly in expressing the notion of temporality through verbs. The present study is an attempt to tackle this issue.
II. BACKGROUND OF THE STUDY

© 2011 ACADEMY PUBLISHER

1550

THEORY AND PRACTICE IN LANGUAGE STUDIES

A. Swain’s Output Hypothesis
As mentioned earlier, Swain (1985) considered an active role for output in L2 learning and believed that output is crucial to push learners to move from semantic processing to syntactic processing. She specified four functions for output: fluency, hypothesis testing, metalinguistic and noticing (Swain, 2005). Fluency function posits that by producing language in a meaningful context, learners can practice accessing L2 knowledge and as a result of recurrent practice, the speed of their access may increase. Another function of output is hypothesis testing function. Swain (2005) claimed that learners sometimes use their output to learn about the target language forms in a trial-and-error fashion. In other words, they use a target form and change it upon receiving feedback from their interlocutors. Modifying their output engages learners in processes believed to be part of the second language learning process (Swain, 2005).
Metalinguistic function holds that we can use language to reflect on language, either produced by ourselves or by others.
This act, swain claims, can mediate second language learning. Producing output ─written or spoken─ also provides learners with opportunities for noticing the problematic areas in their interlanguage. As a result, they may pay attention to the things they have to learn. This raised attention prompts the cognitive processes happening while the learner is learning a new “linguistic knowledge or attempts to consolidate the existing one” (Swain & Lapkin, 1995, p. 384).
B. Output as an Attention-drawing Device
In light of the above theoretical arguments, several studies have empirically examined the role of output tasks in L2 learning (e.g. Colina & Garcia Mayo, 2007; Dekeyser & Sokalski, 1996; Izumi, 2002; Izumi & Bigelow, 2000; Izumi et al., 1999; Izumi & Izumi, 2004; Mennim, 2007; Nassaji & Tian, 2010; Qin, 2008; Reinders, 2009; Song & Suh, 2008;
Swain & Lapkin, 1995; Yoshimura, 2006). One of the first studies of this kind was Swain and Lapkin (1995), which attempted to provide support for the “noticing” function of output. They gave young learners a writing task and taperecorded their speech during task performance to check the cognitive processes involved in L2 learning. They found that young adult L2 learners do notice the gaps in their linguistic knowledge during language production and also engage in cognitive processes such as grammatical analysis believed to be effective for L2 learning. The role of output, especially its noticing function, was explored in other studies, too. Izumi, Bigelow, Fujiwara, and Fearnow (1999), for example, sought to explain whether an output task encourages the noticing of the past hypothetical conditional in
English through two different tasks. In the first stage of the task, the output group had to remember and reconstruct a passage but the input group had to answer true /false questions on the same passage. In the second stage of the task, the output group had to revise what they have reconstructed while the input group had a free-writing activity in which the use of the target form was not required. The result of the study supported the output hypothesis partially in the sense that the difference between the two groups was statistically significant only on posttest of production for the second stage of the task but not for the first one. The researchers believed that this was because in the first stage of the task the output group had to remember and reconstruct the passage at the same time which was of great memory demand and prevented the participants to concentrate on the target form. Izumi and his colleagues concluded that beside considering how the task encourages noticing, we should take into account how it encourages further processing after noticing has taken place. Izumi and Bigelow (2000) conducted a similar study but found no improvement for the experimental group on any of the posttests─ neither on the recognition test nor on the production test. In an attempt to revise the previous studies, Izumi (2002) conducted another study in which he found that the participants who took part in the output treatments showed greater improvement in learning the target form than the learners who received the same input just for comprehension. In the next study, Izumi and Izumi (2004) changed the modality of the output task and investigated the effect of oral output on the acquisition of relative clauses in English. The output group engaged in a picture description task while the non-output group engaged in a picture sequencing task. The two treatment groups were exposed to the same aural input for the same amount of time. Contrary to the expectations, the results illustrated that the output group did not outperform the non-output group while the non-output group demonstrated greater overall language gains. A careful post-hoc reconsideration of the treatment tasks showed that the output task failed to engage learners in the syntactic processing required for triggering L2 learning, whereas the task for the non-output group seemed to encourage form-meaning mapping better.
While most researchers were interested in finding the effects of engaging learners in output tasks on noticing and learning, Yoshimura (2006) examined the effect of foreknowledge of output tasks on noticing and processing of input.
Put it simply, he told the participants they were going to engage in three types of tasks: exact reconstruction in which they had to memorize the text for later reproduction, content reconstruction in which the participants had to retell the content of the text and finally, reading for visualization (control task) during which they had to draw a picture based on the reading text. The interesting point was that, at the end of the study, instead of performing output tasks, learners took a posttest of comprehension and fill in the blanks. Yoshimura expected the participants to notice the gap in their interlanguage through the first two tasks. The analysis of the data suggested that the production groups were more successful in verb uptake than the control group. Yoshimura concluded that noticing of the language can happen not only during or after the output task but also before production of output if the learners consider input processing as a planning stage for subsequent output tasks.
Mennim (2007) also explored long-term effects of noticing on oral output. He reported on the effects of classroom exercises that promote noticing and conscious attention to form. During an academic year, records of the learners‟

© 2011 ACADEMY PUBLISHER

THEORY AND PRACTICE IN LANGUAGE STUDIES

1551

output were carefully examined for noticing and improvement in the use of the noticed forms. Initial analysis of the tracking of two students' noticing and subsequent use of the form revealed that they had much difficulty at the beginning of the year but nine months later, they showed much improvement in using the form accurately. Through this study, Mennim attempted to show how learners' noticing of the word might be related to this improvement.
C. The Effect of Output Task Types
A few studies have also compared the effectiveness of different kinds of output tasks on noticing and learning of target forms. Swain and Lapkin (2001), for example, engaged immersion students in two tasks i.e. a jigsaw task and a dictogloss task. They asked their students to reconstruct a content which was the same for the two tasks. The researchers hypothesized that the dictogloss task would foster greater attention to form. Contrary to their expectations, task differences did not lead to greater attention to linguistic forms. However, dictogloss students were more accurate than jigsaw students. In addition, dictogloss students attended to discourse structure but jigsaw students did not focus on logical and temporal sequencing. Furthermore, dictogloss students produced more complex linguistic structures and vocabulary. Colina and Garcia Mayo (2007) also examined the effectiveness of three types of output tasks: a text reconstruction task, a jigsaw task and a dictogloss task. The study showed the efficiency of all these tasks in drawing learners' attention to form. However, findings of the study seem to indicate that task type controls the nature of attention.
That is, different tasks draw learners' attention to different features of language. For example, the text reconstruction task focused learners' attention especially on determiners, passive voice and verb form while jigsaw task focused their attention on a wide variety of linguistic features. Besides, the findings of their study indicated that the text reconstruction task caused the greatest attention to form. On the other hand, Song and Suh (2008) who compared the effectiveness of a reconstruction task with a picture-cued writing task in noticing and learning of English past counterfactual conditional reported that the efficacy of the two output tasks was the same in the posttest. In line with the study of Song and Suh (2008), Reinders (2009) explored the effects of three types of production activities on uptake and acquisition of negative adverbs in English. He found that the effect of the three activities ─ a dictation, an individual reconstruction and a collaborative reconstruction task─ was almost the same on the acquisition of grammatical items.
The result suggested that all three activities led to uptake. However, the collaborative reconstruction and the dictation activities resulted in greater uptake than the individual reconstruction activity.
While Song and Suh (2008) and Reinders (2009) reported no considerable effect of task type on learning, Nassaji and
Tian (2010) reported significant effect in their study. They compared the efficiency of reconstruction cloze tasks and reconstruction editing tasks in learning English phrasal verbs. Furthermore, they examined the effect of performing the tasks collaboratively versus individually. The results revealed that accuracy of task completion was greater when learners completed the tasks collaboratively comparing with the time they completed them individually. However, collaborative tasks did not result in considerably greater gains of vocabulary knowledge than individual tasks. More importantly, the results suggested an effect of task type as the editing task was more effective than the cloze task in promoting negotiation and learning.
One conclusion that may be drawn from the above studies is that output tasks may facilitate noticing of the target forms, but they may not necessarily lead to better gains of knowledge in comparison to non-output tasks. Nevertheless, as Song and Suh (2008) stressed, studies in this area did not show a clear positive effect for noticing function of output.
Besides, contradictory results were reported on the effect of text reconstruction task on the learning of target forms.
Based on the foregoing discussion, the present study addressed the following research questions:
1. Would there be a difference (or gap) in the performance of the Iranian young adult EFL learners between their comprehension and production of verbal morphemes (present progressive –ing and simple past –ed)?
2. If it is so, would output task (i.e., text reconstruction task) fill the gap between comprehension and production of verbal morphemes (–ing and –ed) by Iranian young adult EFL learners?
3. Would output task affect Iranian young adult EFL learners‟ acquisition of verbal morphemes?
III. METHOD OF THE STUDY
A. Participants
The participants were chosen from among students (within the age range of 13-15) enrolled in the Iran Language
Institute (Babol branch), an important language institute in Iran. The two selected classes were similar in their level and student population. The selection of the participants was based on the results of the pre-test on the target forms. The participants who demonstrated an acceptable knowledge of the target linguistic forms but had yet to improve their knowledge of the morphemes were invited to participate. During the pre-test, those students who showed sufficient knowledge of the aforementioned morphemes (scored 90% or above) and those who did not demonstrate any sign of knowledge of these morphemes both in their comprehension and production were eliminated from the participants‟ pool.
Thirty-nine out of forty two students who participated in this study were assumed to be developmentally ready to learn the target form. All of these students agreed to participate in the study. One group was randomly chosen as the control group and the other one as the experimental group.
B. Procedure

© 2011 ACADEMY PUBLISHER

1552

THEORY AND PRACTICE IN LANGUAGE STUDIES

The procedure of the study was as follows: first, one of the researchers administered the pretest and scored the test papers in each class to screen the participants meeting the requirements of the study. The pretest was conducted in three sessions - one session for comprehension test and the next two sessions for production tests all of which were in written modality. Two production tests were given to the learners to check their production of verbal morphemes both in and out of context. The reason for using two production tests was that the researchers were interested in checking the learners‟ production at sentence and discourse levels. The interval between the pretest and the first practice session was exactly one week. The practice sessions were also in written mode and included three successive sessions with a oneweek interval between them. The students were told that the purpose of the study was to help them make progress in
English writing. Before beginning the first session, the participants were made familiar with the procedures of the practice sessions. After the practice sessions, the posttest was administered in three sessions with the same order as employed in the pretest. There was a one-week interval between the last practice session and the first session of the posttest. C. Instructional Treatment
1. Text reconstruction task
The task used in this study was a modified version of a text-reconstruction task used in previous studies on output.
The whole treatment was done using an audio-text cartoon strip from a series published by Oxford University Press
(Toyama & Rivers, 2005). The input was controlled for length, which helped the researchers have full control over the time of input exposure and succession of the activities which eventually made data collection easier. This cartoon strip was specially selected because it was expected to be reasonably, though not intimately, familiar to the participants. It was also expected that the content would constitute a feasible processing load in terms of understanding for the learners.
After all, the story was suitable for the purpose of the study as it consisted of two parts: present and past. Following
Izumi (2002), the whole text was divided into two parts to lighten the memory load on the learners. Then, the learners had to read and listen to two two-minute audio texts which were semantically related to each other. The pausing point was carefully chosen so that each part remained coherent and meaningful. There were about 20 sentences in each part so it was neither very long for the learners to forget the content of the story nor very short to memorize the sentences.
An important modification in this text reconstruction task was the cartoon- strip accompanying audio-texts. The cartoon strip seemed to lighten the memory load on the part of the subjects when they were reconstructing the texts and to help them concentrate on the form rather than on recalling the story line. Audio text was selected instead of common texts, again to help the learners in reading, and as a result, to provide them with more time and attentional resources to devote to forms. The text was not artificially made for this task. Rather, it was an authentic text to give learners a chance to prepare for confronting with genuine texts in real life situation. Furthermore, the text was carefully selected to have sufficient amount of target forms to attract learners‟ attention.
2. Treatment procedure
At first, the goal of the task was conveyed to the learners in the output group. They were told to a) read and listen to the story carefully and try to understand it and b) reconstruct the text as accurately as possible. Then, the emphasis was on comprehension so that the real aim of the study does not become clear to the participants. Before beginning the task, the participants were told that they have to read and listen to the audio text carefully because it will not be repeated for them before they reconstruct it for the first time. They were instructed not to take any notes during input exposure phase.
They were also informed in advance that they will have limited amount of time for production to account for time variable. After the first input exposure, the learners had to reconstruct the text as accurately as possible. Similar procedure was adopted and employed in the second and third sessions. When the learners reconstructed the whole text, the same procedures were repeated. At this point, the learners were instructed to make any changes they thought necessary to make their written productions as accurate as possible. Again, the learners were not allowed to take notes during input exposure. The same procedures were followed for the second input exposure. No feedback was given before or during the study about tasks or students‟ performance so that only purely task-based effects are gathered.
D. Testing Instruments
To assess learners‟ knowledge of the target forms before and after the treatment, three different testing measures were used. A multiple choice test of verbal morphemes was used to check receptive knowledge of the forms. To check their productive knowledge of the target forms, first, an individual picture description task was used. Next, a picturedescription task in context was employed which included two parts: one aimed at testing learners‟ production of –ing morpheme (present progressive tense) and the other for testing their production of –ed morpheme (past tense). The tests were given in the order mentioned above. Equal number of items was allocated for each of the targeted morphemes in each test. In the context picture-description task, the number of items was also equal but each of them was aimed at specifically testing one of the morphemes. The exact time for each test was determined in advance based on the results attained from pilot studies. To avoid drawbacks of learners‟ fatigue due to taking several tests, the testing sessions were held separately─ one test in each session in the above mentioned order (see Appendices A, B and C).
1. Recognition test
The recognition test consisted of 20 multiple choice items adopted from Bloor et al. (1970) (see Appendix A for sample items of the recognition test). Half of these items were to check learners‟ ability to recognize correct use of

© 2011 ACADEMY PUBLISHER

THEORY AND PRACTICE IN LANGUAGE STUDIES

1553

present progressive –ing and the other half was allocated to simple past –ed. The participants were instructed to choose one answer which is correct. One point was assigned for each correct answer and no point was given when they chose incorrect answers or gave no answer. Then, the total score for all items answered correctly was 20.
2. Individual picture description task
Following Sung and Suh (2008), the individual picture description task was employed in the study (see Appendix B for sample pictures of the individual picture description task). The pictures were adopted from Toyama and Rivers
(2005) published by Oxford University Press. In the present study, this test was used as a written elicitation measure to assess learners‟ productive knowledge of target forms. For each item in this test, learners saw the picture and described it in a simple sentence. In each picture, the time was clearly mentioned. The learners were asked to pay attention to the time mentioned in each picture and try to use it in their sentences. The test included 20 items and the time allocated for each item was one minute. Decisions about the time for each exam were made based on feedback received from the pilot studies.
3. Picture description task in context
The picture description task in context was specially made for this study as the researchers were interested in assessing the learners‟ ability to use verbal morphemes in context where discourse rather than isolated sentences was the unit of producing language. Two pictures were selected for this purpose─ one indicating time in the past and the other one indicating present. The activities the characters were doing in each picture were numbered. Thus, the learners were expected to produce eight sentences for each picture. We thought that the learners might produce more than eight sentences for each picture since they were going to write at discourse level, but in scoring, just those eight items were taken into account for each picture. So, they had to produce at least sixteen items for the two pictures.
E. Scoring and Analysis
All the test data were scored by giving one point to a correct answer and zero for an incorrect answer. The data from each test were scored separately. Owing to differences in the number of items included in each test, scores obtained from each testing measure were changed into percentage. Finally, the composite score was calculated using the results of all three testing measures. It is worth mentioning that the composite score for the two production tests was also obtained to have a clear idea of output tasks on learners‟ productive knowledge. On all production tests, only the production of the targeted morphemes was considered as correct response. The errors related to spelling were ignored since the researchers were only concerned about using correct morphemes for the related tense.
IV. RESULTS
To answer the research questions, the statistical analysis of SPSS was performed to determine 1) whether there is a difference between the learners‟ recognition and production tests of verbal morphemes, 2) whether there is any significant difference between the two groups concerning their performance after the practice sessions and finally, 3) whether there are any significant changes within groups regarding their receptive and productive abilities over time.
Before examining the effect of treatment, the knowledge of the two groups on the past and present progressive morphemes was examined. To do so, the composite test scores of the two groups on pretests of comprehension and production were submitted to independent-samples t-tests. It should be noted that the data from each testing measure were scored separately and were then combined to obtain a composite test score. The result of the analysis illustrated no significant difference on the knowledge of verbal morphemes (their comprehension and production) between the output group (M=54.93, SD=15.96) and the control group (M=48.29, SD=15.13) (df =36, F= .000, p