Realization of preemptive focus on form in the English-language teaching context

While reactive focus on form (FonF) has been addressed extensively in the literature, preemptive FonF has not attracted the attention it deserves. To fill in part of this gap in the English-language teaching (ELT) context, the present study was conducted to examine the occurrence of preemptive focus on form episodes (FFEs) in the classes of two male and two female English-language teachers. Additionally, it aimed to explore the frequency of student-initiated and teacher-initiated FFEs as well as the uptake and no uptake moves in four classes. To address these issues, all cases of preemptive FFEs were identified in 6 h of instruction obtained from videotaping of four classes. Results confirmed that preemptive FonF does occur in the process of meaning-focused communication and that they are used by male and female teachers almost equally to deal with linguistic difficulties. The findings further showed that studentinitiated and teacher-initiated FFEs occurred almost equally in four classes. As to the type of moves, no uptake move was recorded to occur more frequently in male teachers’ classes and to occur more frequently in student-initiated episodes in all classes. Results can raise the awareness of ELT teachers about the benefits of employing preemptive FonF in the context of meaning-focused communications.


Introduction
Discussions of language pedagogy have recognized distinctions between meaning-focused and formfocused instruction (FFI; Ellis et al. 2001b). As Howatt (1984), Skehan (1996) and Willis (1996) maintained, the strong version of communicative language teaching (CLT) favors a meaning-focused instruction which highlights the primacy of communication for acquiring linguistic knowledge. As they maintain, this type of instruction is reluctant to attribute any benefits to the direct instruction of linguistic forms. In contrast to the meaning-oriented types of instruction, FFI advocates efforts, either implicit or explicit, to draw learners' attention to the language (Spada 1997, Ellis et al. 2001b. Though much benefit has been attributed to the meaning-focused instruction, scholars (e.g., Doughty 2001, Ellis et al. 2001b) have criticized a mere attention to meaning through communicative activities for its failure to develop high levels of grammatical knowledge in learners and propose that meaning-focused pedagogy be supplemented with FFI. Long (1991) distinguishes between two types of FFI, namely, focus on forms (FonFs) and focus on form (FonF). According to him, in contrast to the FonFs instruction which involves teaching isolated linguistic forms in lessons of a structural syllabus, FonF instruction "draws students' attention to linguistic elements as they arise incidentally in lessons whose overriding focus is on meaning or communication" (pp. 45-46). Other scholars (e.g., Spada 1997, Ellis et al. 2001b, Ellis 2016) have classified FonF into planned and incidental types. According to Ellis et al. (2001b), in planned FonF, instructors, prior to the class instruction, decide on a form for attention and design communicative tasks to provide opportunities for its practice. Incidental FonF, however, occurs when a particular form or forms stand out while performing communicative tasks. In other words, in this type of FonF, language forms are not preselected, but they attract spontaneous attention of instructors while involved in class routines. The existing research (e.g., Lyster and Ranta 1997, Lyster 1998a, Ellis et al. 2001a, 2001b, Basturkmen et al. 2004, Ellis 2015 has also documented that incidental FonF consists of reactive and preemptive types. Basturkmen et al. (2004) define reactive FonF as responses to errors made by the students, and preemptive FonF as "queries raised by either the teacher or the students about a linguistic item even though no error has occurred" (p. 244).
As Ellis et al. (2001b) maintain, researchers and teacher educators have, primarily, concentrated on reactive FonF as the main discourse mechanism for attention to forms during instruction; however, preemptive FonF has been more of a neglected aspect of classroom teaching. This study was planned to probe into this neglected aspect of English-language teaching (ELT) classroom discourse. This enquiry was prompted by theories of second language acquisition (SLA), which underline the need for attention to form in the context of meaning-focused activities. In particular, this article is situated within the tradition of instructed SLA, which involves incidental and intentional and implicit and explicit L2 learning, when learning processes are intended to be influenced by teachers, classmates or pedagogic materials (Long 2017). We adopted the definition for FonF posed by Ellis (2016) as a central construct in task-based language teaching which refers to specific kinds of "activities" or "procedures" rather than to an "approach". Specifically, this article explored the realization of preemptive FFEs in classes of different gender teachers. To our knowledge, the existing literature has largely overlooked the contribution of social constructs such as gender in the representation of preemptive FonF in ELT classes. In addition to documenting the frequency of preemptive FFEs in classes of male and female instructors, this study touched upon other neglected aspects of ELT classroom interactions, namely, the occurrence of studentversus-teacher-initiated FFEs, the rate of the uptake versus no uptake moves and the aspects of the language FFEs addressed.
It is worth to note that, as maintained by Williams (1999), little FonF occurs in student-student interactions; and in order to deal with the data in detail, we only analyzed teacher-student interactions. Also of equal importance to state is that, due to the significance of uptake move in clarifying learning and the fact that the majority of studies in the field have used uptake as an indication of the effectiveness of incidental FonF (e.g., Lyster and Ranta 1997, Oliver 2000, Ellis et al. 2001a, 2001b, Mackey et al. 2003, Loewen 2004, we committed ourselves to scrutinize this move to document the pattern, if any, of its occurrence in the classes of different gender teachers. The need to investigate gender differences in the use of linguistic forms has been highlighted by scholars like McConnell-Ginet and Eckert (2003). As to the use of FonF, realizing the differences in the use of this construct in the interactions of different gender teachers will help to gain a better understanding of cross-gender variation in the realization of this construct, which, consequently, contributes to more efficient communication between genders. Additionally, due to the benefits attributed to the use of preemptive FonF in language instruction, uncovering its manifestation in classroom interactions can help enrich the literature on how preemptive FonF is realized in real practice. It might also direct the attention of ELT teacher educators toward the inclusion of talks on preemptive FonF in their teacher preparation programs. Results can also raise the awareness of ELT teachers about the benefits of employing preemptive FonF in the context of meaning-focused communications.

Theoretical Investigations
According to Fotos (1998) and Skehan (1998), FFI appeared as a reaction to the strong version of CLT, which focused on purely communication-oriented tasks at the expense of the formal aspects. The rationale behind such a movement was the fact that language learners "typically do not achieve high levels of linguistic competence from entirely meaning-centered instruction" (Ellis et al. 2001b, p. 408). To deal with this dilemma, FFI appeared, so as to put forward an effective way of formal instruction in a communicative classroom (Doughty and Williams 1998).
In addition to the controversy over meaning-or form-centered instruction, debates have been recorded in the tradition of FFI over the degree to which teachers should attract students' attention toward grammar while keeping the communicative nature of the treatment at the required level. This controversy has led to two main positions within the form-focused model, namely, FonFs and FonF. Ellis, Basturkmen and Loewen (2002) define FonFs as the systematic and intensive treatment of linguistic aspects of a language through a preplanned design, which amounts to some separate attention to grammar. FonF, however, is defined as a feature of CLT on the premise that "although CLT is meaning-centered, it needs not preclude attention to form if this takes place in the context of performing a communicative task" (Basturkmen et al. 2004, pp. 243-4). According to Sheen (2002), the focus on the language should be to the extent that no or, at least, minimal interruption occurs in communication. Williams (2001) believes that this practice needs to be brief, unobtrusive and implicit, so as to keep a balance between "bringing the form into the learner's focal attention and allowing the form-meaning mapping process" (p. 326).
FonF has been viewed to be of two distinct types: proactive FonF and reactive FonF or planned (proactive) FonF and incidental (reactive and preemptive) FonF. According to Doughty and Williams (1998), both proactive FonF and reactive FonF occur in communicative contexts. To them, proactive FonF, which can occur either implicitly or explicitly, involves conditions in which a teacher selects a specific form of language in advance, so as to assist students in performing a communicative activity. Reactive FonF, which can also take an explicit or implicit form, arises in a context where class members, during a communicative task, deal with an utterance a student produces erroneously.
Another categorization of the types of FonF distinguishes planned from incidental FonF. According to Ellis et al. (2002), the former involves the use of predetermined, focused communicative tasks to elicit a specific linguistic form from the students in the context of meaning-centered language use. As to incidental FonF, Ellis et al. (2001b) maintain: Incidental focus-on-form involves the use of unfocused tasks, i.e. communicative tasks designed to elicit general samples of the language rather than specific forms. Such tasks can be performed without any attention to form whatsoever. However, it is also possible that the students and the teacher will elect to incidentally attend to various forms while performing the task. In this case, of course, attention to form will be extensive rather than intensive. (p. 421) Two types of incidental FonF, i.e., reactive FonF and preemptive FonF have been distinguished by a number of scholars (e.g., Long and Robinson 1998). Ellis et al. (2001b) maintain that reactive FonF arises when learners produce an utterance containing an actual or perceived error, which is then addressed, ordinarily, by the teacher and, sometimes, by another learner. Thus, it supplies learners with negative evidence. Whereas reactive FonF arises when there is something problematic in learners' utterance and involves negotiation, preemptive FonF is problem oriented and involves "the teacher or the learner initiating attention to form even though no actual problem in production has arisen" (Ellis et al. 2001b, p. 414). According to Ellis et al. (2002), while the former addresses linguistic gaps in a learner's language knowledge, the problem raised may not relate to other students. On the other hand, preemptive FonF instruction can be initiated by the teacher who feels that the problem is that of most of the learners. This approach may be justified on the grounds that students do not simply notice a point which, to the teacher, is very important; accordingly, it is the teacher who should discharge the responsibility of drawing learners' attention to the point, which may involve interruption to the flow of a communicative activity.

Empirical studies
In addition to the theoretical investigations of different types of FonF, empirical studies, mostly of descriptive nature, have explored the occurrence of different types of FonF in the actual practice. As will be evinced in this section, within the existing empirical studies of incidental FFEs, some (e.g., Haque 2016) have compared the appeal of planned and incidental FonF; others (e.g., Eslami andKung 2016, Van de Guchte et al. 2016) inquired into the incidental type in general, not distinguishing between the two types; some publications (e.g., Pouresmaeil and Gholami 2019) compared and contrasted the distribution of reactive and preemptive types of incidental FonF in their data, and others (e.g., Khaki et al. 2015) studied either type of incidental attention to form. The scholarship can be further differentiated in terms of the type/types of interactional patterns they isolate and the details of FFEs they explore.

Planned versus incidental FonF
The dichotomy of planned versus incidental FonF has generated studies of both qualitative and quantitative nature. Within the qualitative tradition, Haque (2016) documented the efficacy of planned and incidental FonF from the viewpoints of scholars, learners and teachers. He found that the educational context, entailing the teaching and learning culture which prevails in the concerned community, is the most important factor in deciding when to choose planned or incidental FonF instruction. Looking at the issue from a quantitative perspective, Shintani (2015) studied 30 Japanese children's learning of plural -s and copula be in two different types of instructions, including incidental FonF and planned FonF. She found that learners in the planned instruction classroom acquired none of the target grammatical features, while children in the incidental classroom showed acquisition of plural -s but not the copula be. The author justified her learners' acquisition of the plural -s on the grounds that they needed to process the grammatical feature for completing a communicative task. For instance, her learners needed to attend to whether the nouns in teacher's commands were plural or singular, so as to acquire the outcome of the tasks.

Incidental FonF
In addition to the studies that explored the contribution of incidental FonF instruction, in contrast to the planned type of FonF, to fostering learners' acquisition of grammatical features, other publications focused on incidental FonF per se, disregarding how it contrasts with the proactive attention to forms. In this literature, we found scholars (e.g., Loewen 2004, Van de Guchte et al. 2016, Eslami and Kung 2016 who addressed the effect of incidental FonF on their participants' language production, mainly in the context of task-based language teaching and the ones (e.g., Williams 1999, Zhao andBitchener 2007) who explored how FonF is realized in the context of various types of interactions in language classes.
Within the literature on learners' production, Loewen (2004) highlighted the primacy of teachers' employing incidental attention to form in the context of New Zealand language school classes. By testing learners' immediate uptake of incidental focused forms (i.e., their incorporating the focused item into an utterance of their own) with tailor-made tests, he found that students scored correctly or partially correct on about 60% of the items during an immediate posttest and 50% on their delayed posttest. Van de Guchte et al. (2016) documented how incidental FonF teaching, with a follow-up task repetition in the post-task phase in the context of task-based language teaching, improved Dutch learners' acquisition of two German grammatical structures. Participants were randomly divided into two groups, which received two subsequent interventions, each of which consisting of a task with a different target structure. During the main tasks, both groups experienced FonF instruction for both target structures. In the post-task phase, however, one group repeated a similar task. Posttests and delayed posttests revealed that the experimental group outperformed the second group on both of the target German grammatical structures in terms of the written accuracy and metalinguistic knowledge. While documenting the primacy of employing incidental FonF instruction in language classes, the authors confirmed the importance of repetition of a similar task in promoting acquisition of grammatical forms. Looking at the issue from a cross-cultural perspective, Eslami and Kung (2016) explored the occurrence of incidental FonF and its effect on the task production of learners of different dyads in an online context. Participants formed 14 nonnative speaker (NNS)-NNS and 16 native speaker (NS)-NNS dyads in an online synchronous computermediated communication (SCMC) context and were asked to complete two multiphase communicative tasks. Language-related episodes (LREs) were used for individualized tailor-made tests to assess students' learning outcomes. Findings demonstrated that in both dyadic types, incidental FonF of the LREs in the SCMC context enhanced students' language development.
As stated above, the literature on FonF has also yielded empirical publications that explored how this approach to instruction is realized in learners' classroom interactions with their peers and their teacher. Williams (1999) examined the occurrence of incidental FonF in learner-learner interactions and found that learners generated FFEs, though not very often (7.77 episodes per 10,000 words). Zhao and Bitchener's (2007) study pointed to the representation of various aspects of FonF in a wider range of interactional patterns, including teacher-learner and learner-learner dialogues. Specifically, they investigated types of FonF, types of feedback, target linguistic forms and types of immediate uptake in 10 h of classroom interactions. Their data analysis showed that, in the two interactional patterns, there existed a significant difference in the types of FFEs. Considering the type of feedback provided, no difference was found between the teacher-or learner-initiated interactions. In terms of the uptake responses, however, a difference was revealed between the two interactional patterns. The authors concluded that in both interactional patterns, incidental FFEs occur frequently and that high frequency of immediate uptakes provides favorable conditions for L2 learning.

Reactive FonF and preemptive FonF
In addition to the studies reported above which did not differentiate the types of incidental FonF, other publications have inspected the distribution of reactive FonF and preemptive FonF in language learning classes. Ellis et al. (2001b), for instance, examined the density of reactive and preemptive FFEs in 12 h of meaning-focused instruction. They found that there were as many preemptive FFEs as reactive FFEs. They further explored the manifestations of preemptive FFEs and found most of them to be based on vocabulary and initiated by learners. Additionally, preemptive FFEs were found to be typically direct, that is, they dealt with form explicitly rather than implicitly. Finally, their students were more willing to uptake form if the FFE was student initiated.
By the same token, Farrokhi and Gholami (2007) investigated how frequently reactive and preemptive LREs were used in 24 h of instruction in one class with two qualified instructors. The study also examined the potential of the types of FonF for leading to uptake. Findings indicated a significant difference in the frequency of reactive and preemptive LREs and between the two types of preemptive episodes. Based on the low number of uptake moves in the findings, a new definition of uptake was suggested, which encapsulates camouflaged uptakes along with learners' immediate responses to FonF. Asadi and Gholami (2014) were, similarly, preoccupied with examining the rate of preemptive and reactive FFEs in language teaching. Their corpus consisted of an 8-hour ELT talk show of 16 male learners of English. Findings revealed that reactive FonF was more frequent than preemptive FonF; and preemptive FFEs were mostly vocabulary oriented, while reactive FFEs were mainly grammar oriented.
Along the same lines of inquiry, Pouresmaeil and Gholami (2019) investigated the occurrence of reactive and preemptive FFEs, their linguistic foci and the rate of their uptake in students' performance. Their data consisted of 15 h of classroom interactions in an upper-intermediate free discussion ELT class with 15 homogeneous adult learners. Their findings revealed that preemptive FFEs occurred very frequently in the class. The highest and the lowest rates of FFEs they recorded were related to lexical items and pronunciation, respectively. Further analysis of their data revealed that vocabulary received the highest rate of uptake and grammar the lowest. Additionally, they found that multiple uptake moves simultaneously occurred for some FFEs with apparently one linguistic focus, and there were some instances of lexical and phonological uptake moves with no apparent FFEs.
On equal terms with the studies reported above, Baleghizadeh (2010) investigated the nature of preemptive and reactive FFEs in 10 h of meaning-focused instruction of an experienced female teacher teaching 11 adult students in a private language institute in Tehran. In contrast to the previous studies, which detailed the occurrence of reactive and preemptive types of FonF in their language classes, Baleghizadeh found that despite the importance of supplementary FFI, FonF was to a large extent ignored by his instructor participant. His participant reasoned that in a CLT context, learners must not be overcorrected and interrupted and that ungrammatical forms of students were mistakes and not errors.
Having a narrower focus than the studies outlined above, Khaki et al. (2015) probed a language learning context of preemptive FonF instruction. They explored whether different learning styles play a role in a grammar classroom taught based on a preemptive FonF instruction. After determining the students' learning styles based on Paragon Learning Style Inventory, all of them were taught a grammatical structure based on FonF instruction. Analysis of a posttest revealed that though some learning styles outperformed others (for instance, extroverts performed better that introverts), preemptive FonF instruction was effective almost for all the students with different learning styles.
More recently, in the context of an aviation English class, Ghaedrahmat et al. (2019) investigated the rate of uptake following learner-initiated and teacher-initiated preemptive FFEs. They observed and audio recorded 20 sessions of the class and employed uptake sheets to count the number of the uptake moves following all instances of teacher-and learner-initiated preemptive FFEs. The audio recorded data were then checked against the data from the uptake sheets. The findings revealed that uptake moves took place more frequently after teacher-initiated preemptive FFEs.
The preceding paragraphs attempted to sketch a picture of how incidental FonF has been touched upon in the literature on FFI. As this section illustrated, a considerable portion of the literature has explored incidental FonF in general, not distinguishing its two types. Other publications have characterized the distribution of reactive and preemptive FonF in language instruction. It seems that in the history of the scholarship on incidental FonF, preemptive FonF has not received the attention it deserves. In consequence, none of the available studies have described the realization of FonF in classes of male and female instructors. That said, we embarked on the present enquiry to enrich our understanding of how preemptive FonF is realized and manifested in English-language classroom practices and interactions. This understanding can make ELT teachers aware of the benefits of the adoption of preemptive FonF in language teaching and can, also, motivate ELT teacher educators to include instruction on preemptive FonF in their teacher preparation programs.

Method
We stated earlier that the purpose of this study was to examine how preemptive FonF is applied in four ELT classrooms taught by two male and two female practitioners using CLT. The research we conducted was descriptive (as elaborated by Seliger and Shohamy 1989), i.e., we only focused on the naturally occurring phenomenon without experimental treatment. This section outlines characteristics of the research participants, the observed course and lesson details, data collection procedure and method of data analysis.

The participants, the course and lesson details
Participants were four ELT teachers, labeled as teachers A, B, C and D in the present study, teaching general English courses in two private language institutes in Tehran. Teachers A and B, identified as group one in this study, were female and teachers C and D, named as group two, were male. Participants' ages ranged between 25 and 28. Their teaching experience varied from 2 to 4 years. All four teachers had completed their BA degree in English literature and had gone through Teacher Training Courses in the same institute. We obtained their consent to use the data collected to write an academic paper about second language learning; however, we did not inform them about the exact focus of the study as we thought that might influence teachers' performance in the class.
Teacher A, who was 25-year-old and the youngest among the four participants, mostly focused on general class discussion and reading comprehension activities. With around 2½ years of teaching experience, she was spending only 8 h a week teaching in the language institute at the time when we recorded her class. Six female learners varying in age between 14 and 17 constituted her class. Teacher B, around 26 years old, mainly included practices on conversation and listening skill in her class session. With about 2 years of pedagogical practice, she was teaching 12 h a week in the language institute in the semester we observed her class. Her students were seven females between 11 and 16 years of age. Teacher C, mostly covered conversation and writing activities in the class session we observed, was 27 years old. Having taught at various ELT institutes for 3½ years, he was teaching 14 h a week when we were collecting the data for this article. His class included five female students between 15 and 18 years of age. Teacher D, the oldest teacher participant, around 28 years old, mainly covered collocations, listening practices and review exercises in his class time. He had 4 years of teaching experience and was teaching 30 h a week at the same institute at the time of data collection. His class population included 13 females aged between 15 and 20 years.
Other points about the classes are worth noting. First, all students in four classes were Iranian English as a foreign language (EFL) learners studying at the intermediate language proficiency level in two private language institutes in Tehran. Next, engagement, study and activation approach was followed by all four teachers who were teaching the same course book, New Interchange (Richards et al. 2013), along with some complementary materials. The classes were further parallel in terms of their content focus, i.e., improving learners' communicative competence. It should be noted that our teachers followed an integrated approach to language teaching. That is, they asked their students to use a combination (or all four) of the language skills within a single activity (or series of activities) to create a situation that was much more similar to the one pupils might come across in the real-world. On this ground, various language skills and language components were the common focus of all four classes.

Procedure and data analysis
To collect data, one teaching session (90 min) of each teacher was videotaped; this way, we obtained a total of 6 h of teaching. Videotaping was done by one of the students of the class using a small digital handycam, so that we could avoid the observers' paradox to the extent possible. This technique provided us with rich data of classroom interactions, including not only teacher-student interactions but also student-student interactions. As already stated, we only analyzed teacher-student interactions in the present article, so as to deal with the data in more detail.
Following the lead of Ellis et al. (2001b), the data were subjected to a detailed analysis, which involved two stages. In the first stage, we identified all the episodes including the incidental FonF. The episodes were all transcribed to analyze them in more detail in the second stage of the analysis. In the second stage, all episodes were scrutinized, so as to identify whether they possess the attributes of reactive or preemptive FonF. Once the frequency of the two types of incidental FFEs were counted, reactive FFEs were left, so that we could analyze, more carefully, episodes containing preemptive FonF. Next we identified student-initiated and teacher-initiated FFEs in the remaining corpus for within-and across-group comparisons. To determine the uptake or no uptake moves of the student-initiated and teacher-initiated episodes and their linguistic aspects, we further scrutinized the data.
In order to elucidate the steps, we analyzed the corpus, and two excerpts of the classes are provided below. A student-initiated exchange is illustrated in the first example that focuses on vocabulary (the meaning of the word niece). A student raises a question, which is termed as trigger by Ellis et al. (2001b) (turn 1). The teachers' response covers several turns (turns 2, 3, 4 and 5). The student's uptake move, namely, their acknowledgments indicating that they have learnt the point or their attempts to incorporate the point learnt in teacher's speech by reformulating it in other ways (Ellis et al. 2001b), occurs in turn 7. 1. S: Excuse me, what does niece mean? 2. T: Anybody? What does niece mean? 3. T: No body? It is a kind of relation. 4. T: hum… 5. T: When you have a sister or brother and he or she has a daughter, she will be your niece. 6. S: mum… 7. S: So, Maryam is my niece.
In the case of teacher-initiated FFEs, teachers sometimes asked linguistic questions from the students and in some other times attracted their students' attention toward specific linguistic forms. In the former, in some cases, the students themselves answered the question; hence, no signal of uptake was heard in students' performance. In other times, uptake was possible, as the teacher himself/herself provided students with the correct response. In the following extract, which illustrates a teacher-initiated FFE, the class is having a pair work on the conversational turns introduced in their book. The teacher notices the word nickname, writes it on the board and triggers a question (turns 1 and 2). As nobody supplies an answer, the teacher, himself, provides it (turn 3); an uptake follows in turn 4. 1. T: Look at the board. What does nickname mean? 2. T: Who has a nickname in this class? 3. T: If your name is Zahra and others call you Zary, this becomes your nickname. 4. S: (laughter, pointing to her classmate) so your nickname is Faty! Finally, to make sure of the reliability of the findings, the coding schemes of both stages of analyses were checked with the participant teachers, which resulted in a high reliability of more than 0.95. This cross-checking helped us most in identifying the aspects of the language the FFEs addressed. Specifically, in cases we were hesitant over labeling the linguistic aspects of the episodes, we asked the participants about the source of the problem. An instance was the perplexity over naming the episode produced by one of the students including the term "thank" instead of "thanked". Though, initially, we identified the episode as dealing with the grammar aspect of the language, later, through checking it with the participant student, we considered it as dealing with the pronunciation aspect. In the next step, to confirm that variations in preemptive FFEs between and across the groups are significant, chi-square analyses were performed on the obtained frequencies.

Results and discussion
It is to be noted that, as stated in the introduction, this study addressed the realization of preemptive FFEs in classes of different gender teachers, the occurrence of student-initiated versus teacher-initiated FFEs, the rate of the uptake versus no uptake moves and the aspects of the language FFEs addressed in the four classes. In this section, we attempt to elucidate these issues. Table 1 indicates that the analysis of the data of the four classes resulted in identifying 122 cases of incidental FonF (71 instances in group one, i.e., female data, and 51 cases in group two, i.e., male corpus) in around 6 h of classroom instruction, that is, one case in every 2.9 min.

The distribution of FFEs in the total corpus
Though at first glance, one can say that FonF is a general occurrence in the observed classes, the overall occurrence of FFEs, in the present study, was not as frequent as it turned out in the literature. Lyster (1998b), for instance, reported 558 FFEs in 18.5 h of immersion instruction, a rate of one FFE in every 1.97 min. Oliver (2000) found 614 teacher responses to erroneous learner turns in four meaning-centered ESL lessons. Though she did not report the duration of the lessons, from the details provided, it seems unlikely that the time of the classes exceeded 12 h (i.e., approximately one FFE per 1.2 min). Ellis et al. (2001b) identified 448 instances of FFEs in 12 h (i.e., one FFE in every 1.6 min) in an ESL context, including language learners from multiple nationalities. And finally, in Farrokhi and Gholami's (2007) study, a total of 641 FFEs were identified in 20 h of meaning-focused lessons; that is, an average of one instance of FFE taking place in every 1.9 min. Overall, it can be inferred that although the number of FFEs in the present study is lower compared with the reported studies, the high occurrence of FonF instances in the present study per se can be an indication of the primacy of attention to form in the context of meaning-oriented activities for adult lower intermediate learners. As Ellis et al. (2001b) maintain, it is quite normal for adult learners to seek opportunities to learn about language forms even in activities that are meaning centered. Celce-Murcia (1991) had also included age and proficiency level in her list of the variables determining the importance of FonF.
Based on the differences recorded in the rate of FFEs obtained in the present study and the ones in the literature, we can claim that the existing discrepancies in the situations under which various studies were conducted might account for the resultant differences. In contrast with the context of the present study, Lyster (1998b), for instance, studied immersion programs. In Ellis et al.'s (2001b) study, learners were highly motivated to enroll in English-medium academic programs. And Farrokhi and Gholami (2007) observed International English Language Testing System's classes. Moreover, differences in the personality of teachers in different cultures and the mixed nationality of the students in Ellis's study might account for the existing differences. Still another reason for the lower number of the occurrences of FonF in our classes might go back to the teachers' main focus on fluency of the learners, rather than the accuracy which necessitates more FonF. One more reason might concern the years of experience of teachers; teachers in the present study did not enjoy a high teaching experience. Mackey et al. (2004) demonstrated that in contrast to the inexperienced ESL teachers, the experienced ones use more incidental FonF techniques and that inexperienced teachers are less likely to deviate from their planned lessons by employing spontaneous learning opportunities mainly in the form of preemptive FFEs. Further, based on Basturkmen et al. (2004), teaching experience plays a critical role in teachers' practice of FonF instruction; that is, experienced teachers are more successful in on-line decision-making required for incidental FonF in the classroom; hence, they know how to obtain learners' attention during a task without interfering with the communicative flow. Considering these points, the fact that teachers in this study did not enjoy a rich professional and academic background in ELT may partially justify the lower frequency of preemptive FFEs. Table 1, further, details that 39, 32, 23 and 28 cases of FFEs occurred in teacher A, B, C and D's classes, respectively. Specifically, a total of 71 FFEs, namely, one case in every 2.53 min was recorded in group one's classes (i.e., females). This number was reduced to 51, that is, one case in every 3.52 min, in group two's classes (i.e., males). The identified cases were further subdivided into reactive and preemptive FFEs. As indicated, 35 cases of preemptive FonF were recorded in female and 25 episodes in male data.

The frequency of FFEs in female and male teachers' classes
Considering the fact that we endeavored to homogenize the classes in terms of their lesson focus, the proficiency level, age and gender of learners and teachers' educational backgrounds, this difference (though only significant at the 0.10 level; χ 2 = 3.28) in the occurrence of FFEs in general (71 in female vs 51 in male classes) and preemptive ones in particular (35 in female vs 25 in male teachers' classes, though not statistically significant) can be attributed to the gender of the instructors.

Reactive versus preemptive FonF
Scrutinizing Table 1, it was further revealed that no significant difference was observed in the occurrence of preemptive and reactive FFEs within each group (35 vs 36 in group one and 25 vs 26 in group two) and in each class in isolation (18 vs 21, 17 vs 15, 13 vs 10 and 26 vs 25 for teacher A, B, C and D's classes, respectively). This finding is similar to Ellis et al.'s (2001b), i.e., in the almost equal rate of reactive and preemptive FFEs. In their study, FFEs were evenly divided between 223 reactive and 225 preemptive types. The frequency of preemptive FFEs was, however, remarkably more than that of reactive FFEs in Farrokhi and Gholami's study (2007). In general, only 168 (26.2%) instances of reactive episodes occurred in their classes, while there were 473 instances (73.8%) of preemptive FFEs. Taken together, the findings of the present study along with those of Ellis et al. (2001b) and Farrokhi and Gholami (2007) can point to the fact that preemptive FonF is a phenomenon of considerable significance, which needs to enjoy, at least, equal importance with that of reactive FonF in pedagogic discussions and studies of SLA.

Different types of preemptive FFEs (teacher-initiated vs student-initiated)
As already stated, preemptive FonF cases were divided into student-initiated and teacher-initiated episodes. Table 2 indicates that there is a slight difference in the occurrence of student-initiated vs teacher-initiated FFEs in male and female teachers' classes. Of the total number of FFEs in male classes, 12 are student initiated, i.e., 48%, vs 16 cases in female classes, i.e., 46%. On the other hand, 13 instances of the total number of FFEs in male classes are teacher initiated, i.e., 52% vs 19 cases in female classes, i.e., 54%. Though at first glance one might conclude that there is a higher percentage of teacher-initiated FFEs in both groups, a closer examination Table 2 reveals a significantly higher number of teacher-initiated FFEs in teacher B's (female) and teacher C's (male) classes (11 cases, i.e., 64.7%, χ 2 = 4.32, p < 0.05, in teacher B's class and 9 cases, 69.2%, χ 2 = 7.22, p < 0.05, in teacher C's class). This is while a higher number of student-initiated FFEs were recorded in teacher A's (female) and teacher D's (male) classes (10 cases, 55.5%, not a statistically significant difference in teacher A's class, and 8 cases, 66.6%, χ 2 = 5.51, p < 0.05 in teacher D's class). Based on this result, it can be proposed that gender cannot be a main factor in the distribution of student-or teacher-initiated FFEs. Though gender might be a contributing factor, one can hypothesize that its effect can be ceased due to other intervening variables. As an example, personality of our instructors might explain the neutral effect of their gender on the occurrence of FFEs. In fact, in several contacts between researchers and teachers, we found that teachers B and C had a more strict personality; this observation might account for their students' lack of courage to initiate questions. Further studies might be needed to control for the effect of personality of instructors through administering a personality questionnaire or to explore its main effect on the realization of preemptive FonF by neutralizing the effect of other intervening variables.

4.5
The role of gender in the frequency of uptake move in preemptive FFEs Table 3 indicates that the uptake moves occurred with a high frequency in both groups (19 cases, i.e., 54.2%, in group one vs 14 cases, i.e. 56%, in group two). This is in line with the findings of Ellis et al. (2001b); albeit, they mentioned a higher level of uptake in their study, i.e., 74% of FFEs in their two classes. Other studies demonstrated varying ranges of utilizing the uptake moves. Lyster and Ranta (1997), for instance, in their observation of four French immersion classes in Canada, found that only 27% of reactive FFEs resulted in successful uptake production. Similarly, Mackey and Philp (1998), in their study of planned FonF, found that only 33% of the recasts was repeated or modified and the rest was followed by topic continuation on the part of learners. Additionally, Oliver (1995), in a study of primary schoolchildren practising an information exchange task, claimed an uptake rate of 35%. Farrokhi and Gholami (2007) found a much lower rate of uptake moves following incidental FonF (only 15.2% of cases), even though the uptake moves were more frequent following reactive episodes than preemptive ones. Of the 168 reactive episodes they identified, 72 (44%) led to learner uptake; whereas only 43 (9%) of 473 preemptive episodes resulted in uptake. These diverse findings may suggest that the teaching and learning context, considerably, affect the benefit of incidental FonF in terms of successful uptake (Ellis et al. 2001a). As Ellis et al. (2001b) also maintain, the high occurrence of uptake move in the present study might be well explained by the observation that students in private language school settings (like the context of the present study) may be more willing to notice linguistic forms and their uptake can be considered as an indication of their noticing. Other factors that might have affected our learners' production of successful uptake in incidental focus of form might be the timing, complexity and type of feedback (as Loewen 2004 has also maintained).
A between-group comparison reveals no significant difference in the occurrence of uptake move between the two groups (56% in group one vs 54.2% in group two). Additionally, a deeper inspection of Table 3 reveals that uptake move occurred more frequently in student-initiated episodes in the two groups. This finding is corroborated in Ellis et al. (2001b) who believed that an uptake move is much more likely to occur in student-initiated FFEs compared with teacher-initiated ones. A closer examination of Table 3 for within-group comparison reveals notable within-group differences in the occurrence of uptake and no uptake moves. In contrast to teacher A and D's classes, the occurrence of uptake moves is considerably lower than no uptake moves in classroom interaction of teacher B and C's classes. As mentioned previously, we found teachers B and C to be more severe in their conduct compared with the other teachers. This observation might be influential in their students' avoidance of revealing their uptake.

Different aspects of the language employed in preemptive FFEs
A further analysis was done on the corpus to identify the types of linguistic items focused on FFEs. It was found that preemptive FFEs in four classes encompass aspects of vocabulary, grammar, spelling and pronunciation. Table 4 summarizes the frequencies of various forms in distinct sections of the data.    Table 4 reveals that the linguistic form most attended to in two groups is vocabulary. Of the total preemptive FFEs in female classes, 16 cases (45.7%) focus on this linguistic aspect; this figure mounts to 56% of the total number of episodes in male classes. Vocabulary is followed by pronunciation (28.6%), grammar (17.4%) and spelling (8.6%) in female classes. In male teachers' interactions with their learners, however, the hierarchy is grammar (32%), pronunciation (8%) and spelling (4%). A chi-square analysis reveals a significant difference in the occurrence of preemptive FFEs concerned with pronunciation between the two groups (χ 2 = 5.79, p < 0.05). The chi-square analysis further shows that the discrepancy in the records of the other three linguistic aspects between the two groups is not significant.
As mentioned above, in both genders' classroom interactions, vocabulary was the linguistic feature most attended to (45.7% and 56%). Comparatively, a little more attention was drawn to vocabulary in male teachers' interactions. A further glance Table 4 reveals no significant difference in the occurrence of this linguistic aspect in teacher-initiated versus student-initiated episodes in each class. This finding is unlike Ellis et al.'s (2001b) who found that their learners were more likely to deal with vocabulary problems.
Regarding the grammar aspect, though a slightly more attention is given to this aspect in studentinitiated episodes in teacher A and D's classes and in teacher-initiated episodes in teacher B and C's classes, this difference is not large enough to be the basis of any claim. In the pronunciation category, again, significant differences were not observed between the two types of preemptive FFEs in each individual class. The picture turned out to be different in case of spelling; more attention was drawn by learners to spelling of the words in teacher C's class and by the teacher in teacher B's class.
Unlike Ellis et al. (2001b), we could not find any FFEs focusing on the discourse aspect of the language. The reason might be related to the proficiency level of our students who were lower intermediate learners and had not, yet, been trained on the discoursal aspects of the language. According to Budak and Reeves (2015), students' abilities and proficiency in the language they are learning affect teachers' thinking about the quantity and quality of grammar instruction. The present findings, however, are similar to those of Ellis et al.'s (2001b) in that vocabulary is the top aspect recorded in our corpus.

Conclusion
This article has explored a relatively ignored aspect of classroom teaching, i.e., preemptive FonF. To date, ELT researchers and teacher educators have mainly examined reactive FonF as the primary discourse mechanism for directing attention to form in the process of meaning-based instruction. Results of this study confirmed that preemptive FonF does occur highly in the context of meaning-focused communication. Based on the results obtained, we wish to argue that preemptive FonF is, at least, as important as reactive FonF.
This study, also, pointed that preemptive FFEs are employed rather similarly by teachers of different genders to deal with linguistic difficulties. In essence, it seems that both male and female teachers are actually advocating the ideas advanced by SLA theories in terms of the routes for L2 acquisition. This research further added to our understanding that gender does not appear to be a determinant factor in the frequency of student-initiated and teacher-initiated FFEs in four classes. The present study further elaborated on the occurrence of uptake move and various linguistic items in the context of preemptive FFEs. Though the frequency of uptake move was recorded to be lower than no uptake move in male teachers' classes, uptake move occurred more frequently in student-initiated episodes in all four classes. Concerning the occurrence of different aspects of language in the identified FFEs, a direct relation between gender and focus on pronunciation was hypothesized.
Overall, through uncovering the realization of FFEs in the classroom interactions of teachers of different genders, this study can point to cross-gender variations in the use of this construct and, consequently, contribute to more efficient communication between genders. By exploring the manifestation of preemptive FonF in classroom interactions, this enquiry can further enrich the literature on how preemptive FonF is realized in real practice. It might also direct the attention of ELT teacher educators toward the inclusion of talks on preemptive FonF in their teacher preparation programs. Moreover, results can raise the awareness of ELT teachers about the benefits of employing preemptive FonF in the context of meaning-focused communications. Finally, by documenting various types of FFEs, the uptake/no uptake moves, and linguistic aspects they address, this article can make ELT teachers more cognizant of these aspects of their classroom teaching and to benefit from this understanding in improving language development of their pupils.
Clearly, a study of 6 h of teaching does not permit generalizations about the realization of preemptive FFEs in ELT contexts. To feel safe to develop certain degrees of generalization to the EFL instruction, it is proposed that more extensive data be collected from a higher number of ELT classes over a longer period of time, which, undoubtedly, results in a richer database and, in consequence, to a more valid interpretation of the findings. A rich corpus, at the same time, provides a more reliable indication of any potential differences in the use of preemptive FFEs that might be attributable to personal variables. It is also suggested that attempts be made in further studies to investigate in what ways and to what extent preemptive FonF contributes to what the whole educational system is all about, namely, student achievement. Hopefully, detailed descriptive information of the type supplied in this manuscript can assist in further research in investigating the contribution of incidental FonF in learners' learning.