1.

Introduction ynch (2001) found students transcribing their own interactions to be an effective way of encouraging learners to reflect on an activity and a means to promote noticing, which is defined as "the intake of grammar as a result of learners paying conscious attention to input" (Batstone 1996). It is generally agreed that noticing is an essential factor in acquisition, and as Schmidt (1990) succinctly states "people learn about the things they attend to and do not learn the things they do not attend to". Having learners attend to their own output, in the form of a transcription of their interaction, so that it essentially acts as input, forms the basis of the learning activity critiqued in this paper.

In many universities in the Middle East, commercial listening text books with native speaker dialogues/monologues are the norm; many teachers and students alike adhere to the native speaker input ideal and generally do not feel they have time and/or do Author: e-mail: [email protected] not see any benefits arising from cooperative feedback sessions involving peers, despite ample evidence to the contrary (for a review see Swain, 2002). For example, Pica, Lincoln-Porter, Paninos, and Linnell (1996) found that NNS-NNS interactions provide learners beneficial feedback as well as comparable levels of input to NS-NNS interactions. Swain suggests that learners should be instructed on how and why to collaborate in order to help encourage what she calls collaborative dialogue, which acts to mediate the acquisition of language by 1. generating new knowledge and/or consolidating existing knowledge (Swain & Lapkin 1994;Swain, 2002), 2. enabling hypothesis testing (Long & Porter, 1985), 3. raising metalinguistic awareness (Selinker, 1972; Larsen-Freeman & Long, 1991;Corder, 1981(in Swain 1995), and 4. providing comprehensible input (Krashen, 1982).

The activity presented in this paper combines both Swain's collaborative dialogue and a transcribing task Lynch (2007) refers to as reprocessing output to produce a uniquely rich learning opportunity. There is substantial evidence that having learners transcribe input can lead to language acquisition. For example, both dictogloss (Swain & Lapkin, 1998) and recorded interviews with native speakers (Clennel, 1999) are established activities in language classrooms. However, Lynch (2007) argues that having learners' transcribe their own output (the aforementioned reprocessing output) is of similar if not greater value in terms of noticing, and that it is a much richer source of relevant material, especially with two learners of roughly similar levels.

Swain (1995) states that output promotes noticing, but that many of these errors, which have the potential to become fossilized if not attended to, are missed. Speaking in particular is a skill for which learners fail to notice errors as there is simply so much to process in real time language use (Skehan & Fortster, 1997). This is true particularly with lower proficiency learners due to their attentional resources being extremely limited, and hence they do not have time to attend to errors as they concentrate on meaning rather than forms. This situation highlights the need for effective feedback. However, this raises the question of whether teachers are capable of noticing these errors and providing feedback in classroom situations?

Studies have shown that real-time feedback on speaking can be ineffective and even demotivating, especially for weaker students; this can be due to teacher inconsistency in giving feedback, time constraints, face saving, and finally teachers simply aren't aware of the feedback options available to them (Van Den Brandon 1995in Kim 2009). This paper addresses some of these issues with teacher feedback, but perhaps more importantly examines the degree to which learners can notice their own gaps. c) Benefits of transcribing Lynch (1997) found that learners were capable of noticing a substantial number of their own errors in speaking and argues that more responsibility needs to be given to learners to do the noticing and hence develop autonomy. He states that the three main benefits built in to noticing exercises involving transcription are: paying attention to normally unnoticed detail, negotiating meaning, and the spoken collaboration between learners (2001). At least in part, another advantage from this type of activity can be ascribed to the repetition that comes with multiple examinations of the transcription (Derakhshesh & Baleghizadeh, 2012). And one final feature to add to the list of ways learning is facilitated is the combination of language focused and meaning focused learning, which has been shown to lead to better results than either kind of learning alone (Ellis 2006).

This trial is an adaptation of a study by Lynch (2001). In his study, adult learners were recorded performing a role-play and then asked to transcribe two minutes of the recording and correct any errors. After this was completed, the teacher went over the corrected transcript with the learners and cleared up any remaining problems. The volunteers involved in this lesson are two personable and motivated 18 year old female English foundation students from Oman. Their English is of a lower intermediate level (overall IELTS 4.5) and they have both received six years of English instruction at public school.

In the following paper the whole trial will be referred to as the 'lesson' which is broken up into tasks (split information and transcribing) and individual stages.

The stages involved in the lesson are listed below:

Stage 1 -The initial stage was a split information find the differences task involving two drawings of a house with supplied vocabulary (see appendix A). The learners were recorded performing this task with a phone.

Stage 2 -The learners transcribed a randomly selected two minute selection of their recording from stage 1 (Transcript 1, see Appendix C). The interaction during the transcription process was recorded but yielded very little usable speech.

Stage 3 -They were then instructed to check their transcript, discuss and correct any errors, and change any parts of it until satisfied with the English. (Transcript 2, Appendix D) This process was recorded and the transcript can be found in Appendix E.

Stage 4 -Once the learners were satisfied with their transcript the teacher checked it with them, changing parts that were linguistically incorrect or expressed in a non-target like way. (Transcript Appendix F)

Stage 5 -The learners were asked their views on the task, this interview was recorded and relevant segments transcribed (Appendix G) While the learners were engaged in the above stages the teacher made general observation notes (Appendix B).

The two minutes of transcription revealed numerous incidences of noticing on behalf of the learners, 12 changes were made (shown below in table 1) and the majority of these changes were for the better (9/12).

The learners in the present study did not make a great deal of changes compared to Lynch's study where the learners made on average 28 changes for the two minute transcription. This could be due to the different levels, (he dealt with more proficient adults), the type of activity (he used role-play), the amount of unknown vocabulary (in this study most of it was known) or the amount of time allowed for proof reading (his learners were able to take the transcripts home overnight to type them on computer).

The more advanced (fluent) of the two learners (learner S1), was responsible for initiating more changes

to equally incorrect form. Output as Input: Facilitating Noticing in Tertiary EFL Learners below in table 2) but of these changes, 25% were unnecessary. On two occasions learner S2 drew attention to an erroneous form but was unable (or simply too slow) to correct it and S1 suggested the change. These results (bearing in mind the small sample size) show that the learners shared the initiative and were able to collaborate to make changes. If a larger sample was taken and the pattern of one sided correction by S1 was found to be significant, then this could have classroom implications for deciding on pairings. Matching learners of the same ability could help avoid more advanced students dominating but even if this is not possible, studies have shown that merely being a participant without actually initiating the changes can result in learning. (Ohta 2000)

The types of correction (table 3 below) show that most changes made by the teacher were in the form of lexical corrections whereas the learner's changes involved either grammar or reformulation (changes to achieve a more precise expression). Editing (the removal of the typical features of natural speech including repetition, false starts etc) did not receive much attention due to the learners ignoring a lot of these features during the transcription phase (stage 1). The fact that the teacher still had a reasonable amount of changes to make (45%) indicates that the learners were stretched to the limit of their linguistic ability during the tasks.

The results from above and observations noted in Appendix B show that during each stage of the lesson the learners seemed to be positively engaged -defined by Lee and Anderson (1993) as a psychological process involving attention, investment, interest, and effort expended in learning. According to Walsh (2002) maximizing engagement is conducive to language acquisition and is therefore an important part of any learning activity. For this transcribing lesson, engagement is due in some degree to the learner generated material which generates positive attitudes towards the lesson, as comments (from appendix G) show below: It is worth noting that after the first transcription when the teacher left the room for a few minutes, the learners started listening to the recording again (without being asked) to make sure it was correctly transcribed. While this level of diligence may not transfer to the classroom, it is a promising sign. e) Do the tasks meet the relevant conditions for learning?

i. Conditions for learning from meaning focused listening

The following five conditions are from Krashen (1982):

1. The input for listening is meaningful For both the find the differences and correcting tasks (stages 1 & 3) the input is meaningful because each learner is supplying the other with input to negotiate and collaboratively complete the tasks. For stages 2 & 4 the focus is not on meaning.

The input for the transcribing was interesting for the most part because it was learner generated and partly due to the novelty value of recording and hearing their own voices.

As shown in table 3, several grammatical feature errors were brought to the attention of the learners but only two unknown words were presented (leash and chimney)

The assist themselves through engaging in interactional modifications to make the input more comprehensible and hence, better suited to their interlocutors IL developmental stage. (Long & Porter, 1985). Other assistance included pictorial and vocabulary support for the find a difference task, and the fact that the ideas and much of the language involved were within the learner's experience.

The input for transcription is learner generated so it is familiar and, as Lynch (2001) states, because the transcription is based on a communicative performance that was already successful it may be less inhibiting for learners to review and improve their output. Long & Porter (1985) mention that group work provides a more supportive and hence less stressful setting than with whole class work. Also the lesson is split up into manageable chunks (stages) so as not to overwhelm the learners. It was thought that the use of voice recorders could be a source of stress but the learners stated they were comfortable with this.

ii. Conditions for learning from meaning focused speaking The following conditions are taken from Nation (1995):

1. The learners have the chance to draw on explicit knowledge in meaning focused use. The presence of the picture and vocabulary for the 'find the difference' activity combined with having a patient and supportive listener (stages 1 & 3) Result in learners learning and using new task vocabulary -an example of drawing on explicit knowledge. 2. The learners have the chance to draw on implicit knowledge in meaning focused use Because the tasks and topic are familiar to the learner they are retrieving implicit knowledge. This allows fluency to develop through the use of features already well known to the learners.

These real operating conditions include time constraints, focusing on the message, competing for the floor and interacting. According to Nation (1995) these conditions are encouraged when learners are deeply involved in a task, as they were in this lesson (see "Quality of engagement" section). Nation sums up the contribution of interaction to these real operating conditions by saying "interaction is an ideal way of developing skill in use as well as learning the conventions of interaction" i.e. the best way to learn to interact is by interacting. 4. The learners are involved in demanding tasks that stretch their vocabulary and grammatical knowledge This task certainly tested the learners communicative ability; as shown in the number of corrected and uncorrected errors produced. (see the 'How much noticing' section) The editing task and the pushed output involved in the split information task both helped make learners aware of gaps in their knowledge. 5. The learners receive feedback about errors.

During their interaction, the learners supplied each other with feedback about the acceptability of their utterances (Long & Porter 1985) as shown below in examples 2-6. This combined with the teacher correction meant learners were receiving immediate and accurate post-task feedback.

From the above conditions we can see stages 1 and 3 meet the conditions for learning from meaning focused speaking and listening and in stages 2, 3 and 4 the conditions are met for language focused learning (this analysis is beyond the scope of this paper)

According to Nation (2007) interaction helps learning by providing plenty of comprehensible input, encouraging pushed output, making learners aware of what they don't know and by helping learners develop the language and strategies needed for interaction. in the transcripts (see appendix) which include clarification requests, repetitions, confirmation checks and comprehension checks. According to Long (1985) these negotiations "allow modification of the interactional structure of conversation? a necessary condition for acquisition". The following examples (3-10) from the learners interaction during the lesson, reveal the many and varied opportunities made available for learning. Confirmation checks-eliciting confirmation that the speaker has correctly heard or understood an utterance. Example 3 S1: yes I have a truck S2: a truck? S1: a truck in the garage (1) the truck is /in the right side\

In the following example the learners fail to clear up a misunderstanding regarding the word 'roof' and S2 accepts S1's mispronunciation, perhaps illustrating a worst case scenario involving an interactional exchange.

S2: okay I have two trucks and there is a two dogs one dog is with a man and the other dog is in front of the house. Do you have them?

S1: I have two dogs one dog is with a lady and the other dog is going to get into the house with a man is it same? Recasts -offer the learner negative implicit feedback, a model and an opportunity to notice a gap in their knowledge. These are rarely used during the interaction but are used by both learners. Example 7 S1: where is your entrance door? S2: middle S1: in the middle (ahhh) S2: yes Mackay, Gass and McDonough (2000) have argued that recasts without interlocutor response might be the least effective form of feedback. And so whether learning is occurring here is debatable. As can be seen from the above example S2 may not realize that S1's recast is in fact a more correct utterance as she doesn't repeat the phrase, although it is thought to be a speculative assumption that improved performance in immediately succeeding utterances can be taken as evidence of learning. (Mitchell & Myles, 2004) It is possible however that learners benefit from negative feedback even if they do not perceive the problem (McDonough, 2005). Language related episodes (LRE's) -talking about or correcting the language produced. These do not occur in the split information task (the focus is almost entirely on meaning) but feature often in the editing phase with several utterances directed at questioning the acceptability of their language.

In the following two examples two alternatives are generated and assessed. LRE's allow the learners to use the language while focusing on form and receive explicit feedback about their utterances, and there is evidence that they can be occasions for language learning (Swain & Lapkin, 1998) Negotiating vocabulary -Of the three words negotiated for meaning (leash, chimney and root) none were successfully negotiated (although root was most likely just a pronunciation error)

More interaction (and hence negotiation) could be included in the lesson by telling the learners not to look at each others paper during the transcription stage (stage 2) and encouraging them to discuss what they hear in order to transcribe correctly. Also, making sure they copy out all the redundant parts of speech from the transcription will promote conscious noticing of such things as discourse markers, hesitation devices etc (Schmidt 1990, in Thornbury & Slade 2006). Learners could then perhaps compare their transcript to one obtained from native speakers doing the same task.

The vocabulary used in this task (see appendix A) was mostly already known to these learners and because of this very few words were negotiated. This is an excellent task for enabling learners to negotiate and gain repetitive and generative exposure to unknown or partly known vocabulary (Nation 2001) and so a significant learning opportunity was not fully utilized.

Because of the difficulty encountered with some of the words (chimney and leash) perhaps simple definitions or example sentences could be supplied to help the learners (this is also generative use of the words). However this would negatively impact on negotiation. Additionally, to enable retrieval after the task, Nation suggests learners could be made to reflect on what vocabulary they learned.

In Lynch's (2001) study, learners were able to take their corrected transcript home over night, word process it (make any wanted changes) and bring it back for teacher feedback/correction. By letting the learners take the transcript home this would encourage revisiting the material and would allow the learners to compare any additional changes they made individually at home. This would also permit them to retain and compare clear examples of their unedited and edited transcripts which they can compare with each other and previous work in order to gauge progress and see if the same problems are occurring.

Although the learners involved in this trial were engaged in the task, the topic selection for the split information activity (a scene of a house) could perhaps be changed to a more interesting picture, perhaps more relevant to these learners. l) Record fluent speakers performing the same task Willis (1996) recommends having students listen to recordings of fluent speakers engaged in the same communicative task they have just performed, and then ask them to point out any similarities and differences they notice. This aids learning in the same ways that recasts do (i.e. provides a target-like model plus the opportunity to notice gaps in their knowledge) and according to Willis, gives learners exposure to "accessible samples of real time talk that is immediately relevant to their learning situation".

A good deal of the benefit from this activity can likely be ascribed to repeated focus on the transcription. In a study with Iranian students Derakhshesh & Baleghizadeh (2012) found that requiring students to repeatedly examine transcriptions and then give an oral presentation incorporating error correction was highly beneficial.

IV.

The transcribing lesson outlined here allowed learners at different stages of development (and thus with different needs) to notice different language features, while developing the same skills. Its modular design and learner generated material make for a well supported engaging experience that successfully combined meaningful communication and a focus on form while keeping the learners engaged and on task for a considerable time. Although it is not easy to enable lower level learners to use L2 to discuss form, this trial has shown that it is not beyond the reach of my students. Raising awareness in teachers and students regarding the benefits of using learner output in lessons is needed in the Omani context. Adapting this lesson to a larger classroom will be a challenge, but judging by this trial, it will be a rewarding one.