Investigating the use of speaking strategies in the performance of two communicative tasks : The importance of communicative goal Miros ł

Although the ability to speak is usually seen as a key manifestation of learners’ ability in the target language they are attempting to master, research has not given ample attention to strategies that can be employed to enhance this skill. In fact, the bulk of such empirical investigations has mainly focused on communication strategies that are reactive devices, predominantly used to overcome difficulties in successful conveyance of meanings and messages, and even this line of inquiry has been neglected in recent years. The study reported in this paper aims to partially bridge this gap by examining the speaking strategies that advanced learners of English used in the performance of two communicative tasks, differing in the extent to which the participants were required to make a contribution to their successful completion. The data collected by means of open-ended questionnaires that were administered immediately after the completion of the two tasks yielded crucial insights into the nature of the speaking strategies and the ways in which the communicative goals inherent in task type influenced the choice of speaking strategies.


Introduction
Even though there are many indices that could be used to describe the mastery of a second or foreign language (L2), most people are likely to associate it with the ability to communicate in a variety of everyday situations, which inevitably entails reliance on the skill of speaking.At the same time, successful development of speaking skills poses a major challenge, particularly in situations in which learners have scant access to the target language (TL) outside the classroom (Majer, 2003;Ortega, 2007;Pawlak, 2004Pawlak, , 2006Pawlak, , 2014)), which, despite all the technological advances, is still the case in most foreign language contexts.The difficulty involved in developing speaking skills in the TL has been stressed by many specialists in the area of second language acquisition.Tarone (2005), for example, states that the ability to produce oral language "is the most complex and difficult to master" (p.485), whereas Burns and Seidelhofer (2010) comment that "learning speaking, whether in a first or other language, involves developing subtle and detailed knowledge about why, how, when to communicate, and complex skills for producing and managing interaction, such as asking a question or obtaining a turn" (p. 197).These concerns are echoed by Kawai (2008), who points out: When the learner is not in the target language environment, it is likely that learning to speak that language will be especially difficult, since learners have minimum exposure to the target language and culture, which is crucial to understanding sociolinguistic traits (such as genre and speech styles), paralinguistic traits (such as pitch, stress, and intonation), nonlinguistic traits (such as gesture and body language), and cultural assumptions in verbal interaction.(Shumin, 2002, pp. 218-219) These challenges of developing the ability to engage in oral language production should hardly come as a surprise when we consider that in order to be able to do so learners are not only required to accumulate different types of TL knowledge (e.g., grammar, vocabulary, multiword units, phonology, pragmatics, genre types and purposes of the act of speaking, characteristics of spoken language), but they also have to deploy these resources in real time under considerable time pressure to attain their communicative goals.These are two interrelated facets which are referred to in the literature in terms of the distinction between language as a system and language in contexts of use (Bygate, 2002), form and meaning (Tarone, 2005), oral repertoires and oral processes (Bygate, 2008), but can also be conceived of in terms of explicit and implicit (highly automatized) knowledge or declarative and procedural knowledge (DeKeyser, 2010(DeKeyser, , 2017;;Ellis, 2009).The degree to which learners can succeed in employing the linguistic resources they have at their disposal in actual communication impinges on their fluency or "the degree to which speech flows, and to what extent the flow is interrupted by pauses, hesitations, false starts, and so on" (Derwing, 2017, p. 246).The obvious hurdles learners have to surmount in producing oral language become even more acutely visible when we consider the dominant models of speech production, both the one initially put forward by Levelt (1989) for speaking in the first language (L1) and its subsequent adaptations to the production of speech in an L2, by, among others, Bygate (2002), de Bot (1992), Izumi (2003), and Kormos (2006).All of them posit that the process of speaking comprises the distinct stages of message conceptualization, where requisite semantic concepts are activated and the choice of language takes place, formulation, where semantic, syntactic and phonological encoding occurs, and a phonological plan is created, and articulation, where the phonetic plan is executed, with the process of monitoring affecting all of the stages as well.
Clearly, L2 learners, who, unsurprisingly, often lack the necessary linguistic and pragmatic resources and have yet to automatize those resources that they do possess, are not in a position to devote their limited attentional capacities to all aspects of the process.This difficulty is further exacerbated by the fact that real-time interaction also imposes a number of other demands, such as listening to interlocutors, drawing on appropriate content knowledge, or simply focusing on the attainment of the intended communicative goal (cf.Derwing, 2017;Kormos, 2006;Muranoi, 2007;Pawlak, 2011;Pickering, 2012).This means that L2 learners are bound to encounter a number of problems when trying to speak, problems that can at least partially be overcome by adept reliance on speaking strategies (SSs), which can be defined as actions and thoughts that learners more or less consciously employ to develop speaking skills and use those skills in real-time communication.Viewed in this way, SSs comprise both largely proactive and mainly reactive strategic devices, or such that are intended to aid the process of gaining greater proficiency in speech production and such that are mainly compensatory in nature, deployed when a communication problem arises (i.e., communication strategies or CSs).However, not only has the bulk of the available research focused predominantly on the latter but, for some reason, the interest in speaking strategies in general has abated in recent years, one manifestation of which was the difficulty in finding a contribution on SSs for this special issue.For example, Cohen and Macaro (2007) elected to include in their landmark edited volume on language learning strategies a chapter on CSs rather than SSs.Additionally, to the best knowledge of the present author, the last major publication on CSs was the one co-edited by Kasper and Kellerman (1997a).
The present paper is aimed to partly fill this gap by reporting the results of a study which investigated the use of SSs in the performance of two communicative tasks by Polish university students majoring in English.The first part will provide a succinct overview of the available studies on speaking strategies, while the second will focus on the aims and design of the study, its findings and the discussion of these findings.The paper will close with a consideration of future research directions into SSs as well as some pedagogic suggestions regarding the teaching of speaking and the role of strategies in this process.

Overview of previous research on speaking strategies
As Pawlak (2011) notes, "for the vast majority of learners, it is the ability to engage in successful oral communication, whether this success is defined as achieving nativelike mastery or merely getting messages across, that drives their motivation to learn a particular foreign language" (p.19).In light of this situation and the formidable challenges that learners are bound to encounter in accomplishing this goal, it is reasonable to assume that skillful use of strategies can on the one hand help learners improve their speaking skills and, on the other, aid them in successfully tackling the difficulties that may emerge in the process of communication.For example, taking as a point of reference the classifications of strategies proposed by Oxford (1990Oxford ( , 2011Oxford ( , 2017)), the use of the metacognitive strategy of planning a speech and the cognitive strategy of practicing that speech several times may lead to better performance in the language classroom, but also outside of the school context, in real-life situations (e.g., a job interview in the TL).The performance can further be enhanced by controlling emotions with the help of metaaffective and affective strategies, ensuring the attention and assistance from the teacher, fellow learners or other interlocutors by means of metasocial or social strategies, or effectively resolving communication problems through reliance on compensatory strategies.Thus, it is both surprising and disconcerting that, as hinted at above, empirical investigations of speaking strategies have somewhat fallen out of favor with specialists, with most of the existing studies zooming in on the compensatory mechanisms that learners resort to when confronted with problems in conveying their messages or successfully participating in interactions.The present section offers a brief overview of such research, first with respect to communication strategies and, second, with regard to more general speaking strategies going beyond the need to deal with immediate exigencies resulting from gaps in different aspects of TL communicative competence.Given the focus of this paper, only the main trends in research on CSs will be highlighted while the studies that have informed the present empirical investigation and those that have examined more broadly conceptualized speaking strategies will be described in somewhat more detail.
Following the definition offered by Faucette (2001), communication strategies can be seen as "the ways in which an individual speaker manages to com-pensate for this gap between what she wishes to communicate and her immediately available linguistic resources" (p. 1).It can thus be assumed that they play a predominantly reactive role by helping learners to deal with problems with getting their messages across in the course of communication.However, Nakatani and Goh (2007), who, incidentally, equate communication strategies with speaking strategies, point out that "there is . . .little agreement about what CSs really are, their transferability from L1 to L2, and whether they can be learnt in the classroom" (p.207).Although Dörnyei and Scott (1997) identified as many as seven possible ways in which CSs can be conceptualized, it is possible to condense them into two major approaches, that is the interactionist perspective, also known as sociolinguistic, and the cognitive perspective, also referred to as psycholinguistic (see e.g., Ellis, 1994;Kasper & Kellerman, 1997b;Nakatani & Goh, 2007).The former views as CSs as external devices learners fall back on in interactions not only to resolve communication breakdowns but also to make communication more effective through the use of negotiation of meaning, selfrepair and time-gaining strategies, which results in the construction of detailed classifications but also underlies the conviction that CSs are teachable.The latter lays emphasis on the mental processes that learners engage in when they experience a language deficit, with the effect that the focus is primarily on compensatory devices, the classifications are thus much more parsimonious and the value of instruction in CSs is denied in line with the belief that such strategies can be transferred from the L1.These two approaches have triggered a spate of research projects, mainly in the 1980s and 1990s, that have aimed, among other things, to identify and describe the CSs used in different situations (e.g., Nakahama, Tyler, & van Lier, 2001), to gauge factors mediating their employment, such as proficiency, task type, cognitive style or willingness to communicate (e.g., Hsieh, 2014;Littlemore, 2001;Mesgarshahr & Abdollahzadeh, 2014;Pawlak, 2009), and to look into the effects of instruction targeting CSs (e.g., Benson, Fischer, Geluso, & Joo, 2013;Nakatani, 2005;Pawlak, 2005;Teng, 2012).
It is fitting to devote a little more space at this juncture to the studies that provided an inspiration or a point of departure for the present investigation, even though it is broader in scope in addressing SSs in general rather than just focusing on CSs, that is, those carried out by Nakatani (2006Nakatani ( , 2010) ) and Pawlak (2015).Nakatani (2006) used responses to an open-ended questionnaire administered to 80 Japanese students to first pilot and then develop the final version of the Oral Communication Strategies Inventory (OCSI) that was used with 400 participants to identify factors underlying CSs use.The instrument is comprised of strategies employed to tackle problems in conversation, both in speaking (i.e., socio-affective, fluency-oriented, negotiation for meaning, accuracy-oriented, nonverbal, message reduction and alteration, message abandonment, and attempts to think in English strategies) and in listening (i.e., negotiation for meaning, fluency-maintaining, scanning, getting the gist, non-verbal, less active listener and word-oriented CSs).The tool was correlated with Oxford's (1990) Strategy Inventory for Language Learning and was subsequently filled out by higher-and lower-proficiency students on completion of a conversation task.The former reported more frequent use of some categories of strategies, particularly of the achievement type (i.e., involving the use of an alternative plan to attain the communicative goal).The instrument was then applied, alongside transcripts of recordings and retrospective protocols, to determine the effects of training 62 Japanese learners in the use of CSs on their performance in a conversation task (Nakatani, 2010).It was found that negotiation for meaning and discourse maintaining strategies contributed to the participants' communicative ability but also that there were few instances of modified output (i.e., changes in the initial utterance in response to a signal on the part of the interlocutor).The OCSI was also applied by Pawlak (2015) in his investigation of the use of CSs by 64 English majors in Poland as they were working on two types of communication tasks, which differed with respect to the requirement to exchange information.He found, among other things, that the CSs for speaking and listening were used with comparable frequency and that the participants mainly opted for strategies based on TL and those helping them get messages across, and that greater attainment translated into more frequent reliance on accuracy-oriented and getting the gist strategies.However, the analysis of the data led Pawlak (2015) to identify a number of flaws of the instrument such as its unsuitability to examining CSs use by more advanced learners, excessive focus on detail that may be difficult to understand for respondents, and lack of sufficient emphasis on strategies included in most classifications of CSs, such as circumlocution, approximation or appeal for assistance.The study reported in the present paper is an extension of this investigation, but it relies on qualitative data and places store by the effect on task type on the use of CSs.
Perhaps because researchers have been primarily concerned with different aspects of CSs, relatively little attention has been given to investigating more broadly conceptualized strategies that can be applied to develop speaking skills and enhance speaking performance.In effect, the available empirical evidence is scant, unfocused, fragmented, and therefore exceedingly difficult to synthesize, with major overviews of language learning strategies treating the terms speaking strategies and communication strategies as synonymous (e.g., Cohen, 2012Cohen, , 2014;;Oxford, 2011) or making references to studies in other areas, such as pragmatics.One relevant study, undertaken by Huang and van Naerssen (1987), examined the link between performance on an oral task and the use of language learning strategies by 40 English majors in an institution of tertiary education in China.Interviews conducted with 10 high-and 9 low-achievers revealed that the former were more likely to engage in functional practice, as exemplified by such strategies as actually using the TL for communication, thinking in that language, talking to oneself, or reading in order to obtain the models for speaking.In another study, whose focus, however, was primarily on pragmatics, Cohen and Olshtain (1993) examined the performance of the speech acts of apologies, complains and requests in role-play activities by 15 advanced learners of English as a foreign language.The analysis of the data obtained by means of think-aloud protocols resulted in the identification of four major strategies (i.e., planning, thinking in two or three languages, searching for TL forms in various ways, and paying scant attention to grammar and pronunciation), allowing categorization of the participants into metacognizers, avoiders and pragmatists.Carson and Longhini (2002), in turn, reported a study in which the first researcher kept a diary for eight weeks including in it comments on the process of learning Spanish in Argentina.The analysis of the 32 entries that Carson made with the help of Oxford's (1990) classification demonstrated that she mainly relied on indirect strategies that were in synch with her learning styles, with metacognitive strategies being prevalent.Kawai (2008) describes an informal study in which two proficient adult learners of English were asked to identify in an open-ended questionnaire the strategies that they used for in-class discussions.Both of them pointed to the role of adequate planning and preparation, practicing speaking on a daily basis, starting discussions with their peers and relying on stop-gap strategies when communication breakdowns occurred.Worth mentioning as well is the study by Pietrzykowska (2014), who correlated the data obtained by means of the SILL with the results of the oral component of the end-of-the-year examination taken by 80 English majors.The results were mixed but suggested a positive contribution of cognitive and compensation strategies to the development of speaking skills.
There are also studies that have looked into the effects of strategies-based instruction on the speaking ability in the TL.One of the first, undertaken by O' Malley and Chamot (1990), showed beneficial results of training high school learners of English in the use of metacognitive, cognitive and socio-affective strategies on their speaking performance, as measured by pre-and posttests.Varela (1999), in turn, conducted a study which investigated the impact of strategy training on the development of oral presentation skills of 41 learners of English as a second language in the USA.The intervention focusing on the strategies of grouping, selective attention, cooperation, self-talk, note-taking and self-assessment led to improvement of oral performance and increased reliance on strategies.Moreover, a positive correlation was revealed between the use of strategies and attainment in the experimental group.In yet another empirical investigation involving three intact classes of Spanish, Naughton (2006) examined the effects of instruction in cooperative strategies on the patterns of interaction in the performance of a small group discussion task.The video-recordings of the interactions of triads that were selected in experimental and control groups before and after the intervention demonstrated that the program resulted in the participants engaging more often in the kinds of interaction that aid successful language acquisition (e.g., asking follow-up questions, requesting classifications and providing them, asking for and offering assistance).Also of relevance here is the action-research project reported by Kawai (2008) in which 50 students of engineering in Japan attending a course in English were provided with taskbased strategy instruction aimed to assist their participation in face-to-face and online discussions.The analysis of questionnaire responses prior to and after the intervention resulted in more active involvement in interactions in both environments, but the effect was more pronounced in chat discussions.
As can be seen from the above overview, most of the research on speaking strategies to date has focused on CSs which have been investigated in a systematic manner, although they have evidently fallen out of favor with researchers in recent years.As regards research on more broadly conceptualized SS that are also more proactive in nature, the empirical evidence is quite limited, pertinent studies represent a rather mixed bag, and the main focus has been on the contribution of strategies in general rather than their role in effective use of the TL in specific tasks.The study reported below is intended to partially fill this gap by exploring the use of speaking strategies in the performance of two types of communication-based activities.

Research questions
The present study constituted an extension on the research project conducted by Pawlak (2015), and, using very similar methodology, aimed to investigate speaking strategies that were employed by advanced learners of English in the preparation for, execution and on completion of two tasks which differed in regard to the existence of an information gap.In effect, the following research questions were addressed: 1. What speaking strategies do advanced learners of English use before, during and after the performance of two communication-based tasks? 2. Are there differences in this respect between the tasks with their diverse communicative goals?

Participants
The participants were 20 students, 11 females and 9 males majoring in English, enrolled in the first year of a two-year MA program in a regional Polish university which built upon a BA program in English philology that most of the students had completed in the same institution.At the time of the study, they had been learning English for an average of 12.8 years, with the minimum of 8 and the maximum of 15.5 years.The participants' proficiency could be described as falling somewhere in between B2 and C1 according to the Common European Framework of Reference for Languages, although there was considerable individual variation.The students self-rated their overall mastery of English as 4.25 on a 6-point scale (1 -lowest and 6 -highest), being more confident about their listening rather than speaking skills, as evident in the mean self-evaluations of 4.44 and 3.96, respectively, on the same scale.The mean end-of-semester grade in the conversation class they were attending was somewhat comparable as it equaled 3.41 on a 5-point scale (2 -lowest and 5 -highest), which is typically used for the purposes of assessment in Polish universities.At the same time, perhaps being aware of their lacking abilities in this respect, the participants seemed to be fully aware of the importance of speaking skills in the overall mastery of the TL because they evaluated it as 5.15 on a 6-point scale (1 -lowest and 6 -highest).
Although the main aim of the program the students attended was to lead them to an MA degree in English philology in the areas of foreign language teaching and translation, in addition to seminars and supplementary seminars, they had the benefit of an intensive course in the TL including separate classes in grammar, academic writing, conversation or translation.They also had to attend a number of content classes in literature, linguistics and culture, as well as several electives, and complete a course in German as an additional (third) language.The students varied tremendously with respect to their everyday contact with English outside of the program offered by the university.Some of them travelled to English-speaking countries, regularly communicated with their friends from abroad in English or had jobs that required its use, whereas in the case of others, access to the TL was scant, it was confined to television or electronic media, and rarely did it involve face-to-face interactions.

Procedures, data collection and analysis
In order to collect the requisite data that would go beyond the information that can be obtained by means of questionnaire items divorced from concrete instances of CS use, the students were requested to perform two communicative tasks in pairs.Task One was a discussion activity in which the participants were asked to express their views on such issues as the advantages and disadvantages of living in the country, the pluses and minuses of being brought up in a big family, and the advantages and disadvantages of possessing a mobile phone.Since the students had access to the same questions and not all of them had to participate equally (or at all) to accomplish the communicative goal, the task could be described as optional information-exchange in nature.In Task Two, the same dyads were instructed to identify twenty differences between two pictures presenting a scene in a park.Since the students were not allowed to look at each other's pictures and there had to be a high degree of interaction between them for the communicative aim to be attained, the activity represented a required information-exchange task.
The participants were given about 15 minutes for the completion of the tasks, which was sufficient for the first task but turned out to be inadequate for the second since most of the pairs failed to pinpoint all the differences during that time.Immediately on completion of each task, the students were given the same questionnaire which was intended to obtain data on the use of SSs drawn on in the completion of each activity and can be regarded as a form of immediate report.It included six open-ended items in which they were asked to report on the following issues: (1) strategies used to prepare for the task, (2) strategies applied during the task, (3) strategies used after the task, ( 4) what they paid attention to when performing the task, (5) the processes engaged in when listening to the interlocutor, and (6) the things that were the most helpful during the interaction.In order to ward off any potential misunderstandings and ensure that the students would write what they really thought rather what they merely could express owing to their limited TL resources, the questions were formulated in their L1, and they could also respond in Polish or English, as they saw fit.The participants were also asked to fill out the OCSI (Nakatani, 2006, see above), and the interactions of all dyads were audio-recorded by means of Dictaphones and later transcribed.Such data, however, will not be taken into consideration in the present study as they mainly provide insights into CSs functioning as reactive measures deployed to deal with communication breakdowns.
The decision to focus exclusively on responses to the open-ended items also determined the type of analysis, which was entirely qualitative.It consisted in going over the responses many times in order to identify predominant types of SSs used, adopting as a point of reference Oxford's (1990) classification, as well as pinpointing specific instances of strategic devices employed in the different phases of the two tasks.The detected strategies were also compared as a function of the two tasks.When doubts arose concerning the categorization of a specific strategy, the researcher consulted these cases with a colleague who had considerable experience in investigating the use of language learning strategies.
On the whole, while the study can be viewed as a follow-up on the investigation undertaken by Pawlak (2015) and mirrors its overall design, there were differences with respect to the tasks used (i.e., a different set of questions, a picture description task instead of a jigsaw task), the conditions under which the activities were performed (i.e., pairs instead of small groups), and the focus of the analysis (i.e., open-ended items instead of the OCSI).

Findings
The results are presented here in the order in which the queries appeared in the questionnaire, first in general and then separately for the two tasks, with all the themes touched on being brought together in the discussion section in order to address the research questions posed.
When it comes to the preparation for the two tasks, somewhat unsurprisingly, one can see the predominance of different types of metacognitive strategies, which, however, were often not related to language learning or use but, rather, to the content of the messages that needed to be communicated.It is also worth noting that, in both cases, the participants sometimes mentioned strategies that did not seem relevant and were listed only because the students might have been familiarized with them in their methodology classes (e.g., metacognitive, association, immediate recall, semantic mapping).As to the differences between the two tasks, the optional information exchange activity involved much more focus on content-related issues, such as seeking arguments to be garnered in support of one's stance.As for specific strategies, the participants pointed to brainstorming ideas with their peers, which can be at the same time regarded as the social strategy of cooperation, the metacognitive strategies of planning, paying attention, or overviewing and linking with already known material, and even the cognitive strategies of creating structure for input and output (i.e., summarizing), note-taking or practicing on condition that the interactions were conducted in the TL.As regards the required information-exchange task, there was visibly much more focus on individual work, analysis of the details in the pictures and language-related issues, as exemplified by the search for appropriate vocabulary.In doing so, the students fell back on the metacognitive strategies of organizing, identifying the purpose of the task, and planning for it, the cognitive strategy of practicing, when they repeated the requisite lexical items that might come in handy in the performance of the task, but also the social strategy of cooperating with peers when deciding how to approach the activity.The excerpts below illustrate some of these points (S stands for student and T for task in all cases): 1 1 All the examples were translated into English by the present author.
I considered the questions, focused on the content, and tried to find relevant arguments.The strategies were reading the instructions carefully, mapping, brainstorming.(S6, T1) I tried to recall the vocabulary needed to describe this particular picture.(S6, T2) Brainstorming, mapping; depending on the difficulty of particular things, writing them down on a piece of paper or practicing them in the head.(S11, T1) I carefully analyzed all the details of the picture and tried to organize my thoughts in such a way that the description would be the most accurate.(S16, T2) When the task was actually being performed, irrespective of its type, the students constantly engaged in monitoring and were mainly concerned with getting their messages across in the face of sometimes lacking TL resources.This led to the use of compensatory strategies, which were reactive in character and in fact represented different types of CSs, such as circumlocution or approximation (i.e., findings synonyms that were more or less suitable).In some cases, the strategies even took the form of gesticulation and, much more rarely, reliance on the L1.The main difference between the activities, though, lay not only in the fact that the picture-description task called for the employment of considerably more specific language but it also required considerably greater cooperation between interlocutors.In other words, while the students could focus on their own points of view and arguments in Task One, paying rather scant attention to what their interlocutors were saying, the situation was dramatically different in Task Two, where lack of collaboration and attention to the interlocutor's description could have resulted in a failure to find the differences.As was the case with the preparation stage, there were also answers which were totally unrelated to the questions in hand and only provided evidence for the participants trying to display their knowledge of methodology rather than reflecting on the task being performed.Some illustrative examples are provided below: I asked questions about the details of the picture, which was very helpful in finding the differences.(S16, T2)

I used circumlocution . . . Sometimes I used Polish when I did not know the words I wanted to use. (S15, T1)
I brainstormed all the time.I also used synonyms when I could not find the right words.(S14, T1)

I focused on logical construction of arguments. I used circumlocution and explanation when I wanted to use a specific word or phrase. (S13, T1)
Choosing the right vocabulary, directing the interlocutor to a specific part of the picture that I wanted to describe.(S13, T2) When I did not know a particular word, I tried to explain its meaning.I tried to be systematic in the description of the picture.I asked my partner to describe the place in his picture that I have just described.(S6, T2)

I used mental maps and brainstorming. (S3, T1)
The question regarding the strategies that the participants used after the completion of the tasks proved to be exceedingly difficult, with the students leaving empty spaces or stating that they did not use any strategies at all.There were also many responses which are very difficult to understand given the focus of the tasks because some students mentioned the need to remember the arguments used or the differences listed, even though no such need was signaled in the instructions for both tasks.Although the metacognitive strategy of selfevaluation was present in both activities, it was much more pronounced in the find-the-difference activity (Task Two), with the participants comparing their pictures, identifying the differences, as well as gauging the level of understanding, and assessing the lexis they had used to convey their messages.In addition, in this case, there was evidence for much more reliance on social strategies in the form of cooperation or asking for clarification and verification, with some of those constituting an inherent part of the process of self-evaluation.It would thus appear that a tangible effect of task performance was one of the key factors that turned students into active users of speaking strategies.The following excerpts illustrate some of these points: When asked what they were paying attention to when performing the tasks, most of the students pointed to TL accuracy with respect to lexis, grammar and pronunciation, and speaking in such a way so as to be understood by their partners.Thus, they could be said to have primarily drawn upon the speaking strategy of language monitoring.The conveyance of messages was also emphasized but it was somewhat relegated to the back seat in the case of most of the participants.Generally speaking, it was Task Two, which was clearly more cognitively demanding, that triggered greater focus on content, and this was to some extent necessitated by the need to precisely pin down the differences between the two pictures.Some of these themes are evident in the excerpts provided below: When performing the task, I tried to produce sentences that were grammatically correct.I was also trying to choose the right vocabulary and to make sure that my interlocutor can understand me.(S2, T1) I mainly tried to speak in an understandable way, choose correct vocabulary, and employ grammatically correct language.(S2, T2) I paid attention to what I was saying and the way in which I was speaking so that my interlocutor could understand me.(S3, T1) I paid attention to precisely formulating sentences and questions.(S3, T2) I concentrated on phonetic accuracy, and careful and meticulous description to catch all the details.I tried to be grammatically correct and to use appropriate vocabulary.(S6, T2) I paid attention to appropriate vocabulary, directing the interlocutor to the right section of the picture, and remembering the differences.(S8, T2) My main focus were pronunciation, grammar and vocabulary, the contents of my talk and preparing the arguments.(S10, T1) I mainly concentered on the details of the picture to which my interlocutor was paying attention.(S12 T2).
I paid attention to the phrases describing the location of the objects in the picture so as not to mislead the interlocutor and to describe the objects precisely.(S15, T2) In response to the query about what they were doing when listening to their interlocutors, most of the students pointed to comprehending the messages being conveyed, regardless of the task being performed.This involved reliance on the metacognitive strategy of monitoring, which was often adroitly combined with the social strategy of seeking clarification.In other words, the focus was mainly on content rather than accuracy of what the interlocutor was saying, with negotiation of meaning evidently taking precedence over negotiation of form (Suzuki, 2018).What should be noted though is that participants paid much more attention to what their interlocutors were saying in Task Two rather than Task One, since this was indispensable for the identification of differences between the pictures.Such tendencies are exemplified by the following excerpts: I was trying to understand and to respond in an appropriate way as well as to justify my responses.(S5, T1) I was trying to remember the advantages and disadvantages mentioned by the interlocutor, so I could respond to them.(S6, T1) I was examining my picture in search of the differences and I was trying to locate the parts of it that my partner mentioned.(S6, T2) I tried to interrupt and asked for the same thing to seek confirmation.(S9, T2) I was trying to locate the objects the interlocutor was talking about.(S18, T2) Finally, when asked what helped them the most when performing the tasks, the students most frequently indicated the qualities of their interlocutors, followed by their own attributes and abilities.As to the former, the participants most often pointed to their partner's gregariousness, the fact that they knew him or her pretty well and were therefore less afraid to make mistakes, his or her TL proficiency or conversation skills (e.g., the capacity to refrain from interruptions or to ask appropriate questions).When it comes to the latter, the most crucial was the mastery of English, familiarity with requisite vocabulary, ample knowledge of the topic that may have been covered in class, creativity, or skills in managing the interaction.It was also the only time when the students mentioned affective issues such anxiety accompanying the act of speaking.In a word, it was in the main social strategies that played the most important role but some weight was also given to affective strategies.Importantly, little difference was revealed between the discussion triggered by the preset questions and the picture description task.Tendencies of this kind can be discerned in the following responses: What helped me was knowing the interlocutor well as well as having planning time to think about my views and the topics which were familiar as they had been discussed in class before.(S6, T1) What mattered was the ease with which the interlocutor was describing the picture but also my own knowledge of the needed vocabulary.(S9, T2) Creativity helped me the most as well as the fact that I am not afraid to speak.I have no inhibitions in this respect.(S14, T1).
Adequate knowledge of grammar and vocabulary was key to performing the task.(S 16, T1)

Discussion
At this point, an attempt will be made to address the research questions posed earlier in the paper which concerned the speaking strategies used to perform the communicative tasks and the impact of task type and communicative goal on the choice of these SSs.With respect to the first research question, the analysis provided evidence for the predominance of metacognitive strategies in all the stages of task performance under investigation, that is preparation, execution and follow-up.To be more specific, the participants most frequently engaged in planning their contribution, both in terms of content (e.g., searching out the relevant arguments of examining the particulars of the pictures being compared) and language (e.g., the selection of appropriate vocabulary), monitoring their performance as the task was being executed, in particular with respect to accurate use of TL subsystems (i.e., grammar, vocabulary, pronunciation), and paying attention to what their partner was saying, as well as self-evaluating their success on the completion of the tasks.There was also clear evidence for frequent deployment of social strategies at all stages of task performance (e.g., cooperation, asking for clarification or verification) as well as compensatory strategies, or simply CSs (e.g., circumlocution, approximation, gesticulation, reliance on the mother tongue), as the task was being performed.Other types of strategic devices were used very rarely, and while this may not be surprising in the case of memory or cognitive strategies that may not have suited the nature of the two tasks, marginal reliance on affective strategies was somewhat unexpected.After all, speaking tasks are by nature anxiety-provoking (Gregersen & MacIntyre, 2015) and it would have been natural for the students to draw on strategies that would have allowed them to combat such apprehension.It should also be noted that, as indicated by Oxford (2017) and empirically verified by Cohen and Wang (2018) in the case of vocabulary tasks, the strategies the students employed served several functions at the same time, a good case in point being brainstorming which can be viewed as an example of a metacognitive, social or cognitive strategy.Another important point is that, on the whole, the metacognitive strategy of monitoring was considerably more focused on language-related issues (i.e., grammar, vocabulary, pronunciation) than the messages conveyed when a particular student was speaking, but the reverse was the case when he or she was listening to a partner.What is also striking is the evident lack of focus on pragmatic considerations, which figured prominently in the research project undertaken by Cohen and Olshtain (1993) referred to above, a situation that can perhaps in part be attributed to different types of tasks used in both studies (i.e., role-plays designed to elicit concrete speech acts vs. more general communication tasks).Finally, particularly disconcerting is the fact that the participants often left blank spaces, in most cases provided rather general responses failing to give details of SSs use, or offered answers that were entirely irrelevant.This testifies to the limited knowledge of the English majors in this respect despite the training in L2 teaching methodology they might have received and thus to the need to raise their awareness of language learning strategies in general and speaking strategies in particular.
When it comes to the second research question, the analysis demonstrated that the use of speaking strategies was at least to some extent conditioned by the type of task being performed and the communicative goals the participants were expected to achieve.The differences between the optional and required information-exchange tasks can be summarized as follows: (1) more focus on content, cooperation and content-related issues in the preparation for Task One than Task Two, (2) focus on more specific language (e.g., lexical items) in Task Two as well as considerably more attention to what the interlocutors were saying, as manifested in intent listening but also more instances of collaboration which was necessary to establish the differences between the pictures, (3) considerably more emphasis in Task Two on self-evaluation, both with respect to content and language, and cooperation, on completion of the activity, and (4) greater emphasis on content in Task Two, which can be viewed as more demanding in terms of the use of the available cognitive resources.The existence of these divergences may indicate that the different communicative goals inherent in different communication tasks can be beneficially harnessed in the process of teaching speaking as well as conducting strategies-based instruction in this area, a point that will be taken up in the concluding section.
An undeniable strength of the present study is that it investigated the application of SSs in regard to the performance of specific communication tasks, something that has been rarely done by researchers, thereby providing valuable insights in this respect.However, the research project is also afflicted by some weaknesses.First, the investigation relied on a single data collection tool, that is immediate report, which severely limits insights that can be gleaned from the analysis.Even though this was to some extent warranted by the aims of the study, which focused on speaking strategies rather than solely communication strategies, thus making the use of transcripts or the OCSI superfluous, there can be little doubt that the inclusion of post-task interviews or the administration of a tool specifically designed to tap the use of SSs could have considerably enriched the findings.Second, the number of participants was relatively small, which severely constrains the generalizability of the results, although it should be kept in mind that large number of subjects in qualitative studies may often blur the picture and prevent researchers from detecting clear-cut patterns (cf.Pawlak, 2018).Third, the use of SSs could have been impacted by a host of individual difference variables, such as gender, proficiency, working memory, motivation, learning style, or willingness to communicate, to name but a few.Fourth, it is conceivable that at least some of the participants were not familiar with the concept of speaking strategy, which would account for their indexical responses or even lack thereof in some cases.Such issues should surely be taken heed of in future empirical investigations dealing with task-based application of speaking strategies.

Conclusions, directions for future research and pedagogical implications
The present study undoubtedly constitutes a valuable addition to research on speaking strategies, a field that has been dominated by a focus on largely reactive communication strategies and has given little attention to more broadly conceptualized, proactive strategies which can be drawn on to enhance the process of developing speaking skills and actual engagement in oral interactions.What is more, in contrast to most available studies, an attempt was made to link the use of SSs to the performance of a specific tasks which differed with respect to the communicative goals that the participants were expected to accomplish (i.e., exchanging opinions and views vs. reaching a consensus regarding specific differences).Despite all the shortcomings that the study may suffer from, it did provide interesting insights into the SSs that advanced learners employ when preparing for, conducting and reflecting on a communicative task.Additionally, it provided evidence that the employment of SSs is bound to be conditioned by the type of activity, the demands it places on interlocutors, and the communicative goals it sets.
This having been said, there is clearly a need for further research into the use of strategies in the performance of a variety of language learning tasks that are more or less communicative in nature.Such research should employ multiple data collection instruments, factor in the influence of individual difference variables and look into the influence of other task-related variables, such as the amount of planning.What is of particular importance, given the close link between speaking and listening, studies of this kind should most profitably address the SSs with respect to both skills, thereby following Nakatani (2006), who developed the OCSI, but simultaneously moving beyond problem-oriented, largely reactive CSs to include more proactive speaking strategies.The findings also offer an important lesson for L2 instruction.It is clear, for example, that different types of communicative tasks can be employed to hone different types of skills and abilities, necessitating diverse foci on meaning versus form, accuracy versus fluency, cooperation versus individual work, and the like.Some of the students' inability to name the strategies they fell back upon when executing the two tasks or the general nature of their comments also speak to the necessity of instruction in this domain, with different types of tasks catering to the use of various kinds of SSs.One way or another, it is clear that if the aim of foreign or second language instruction is the development of speaking skills, students have to be requested to complete well-designed communication tasks on a regular basis.This condition, however, is seldom met even in language courses designed for English majors who are expected to become highly proficient in their use of TL in different contexts.

I
tried to write down and remember the arguments used by my interlocutor and the way in which he or she was speaking (metacognitive).(S3, T1) Consulting accuracy with the partner.(S5, T1) Comparing the pictures and differences.(S5, T2) I self-evaluated my performance and compared my score with others from the group.(S6, T2) Checking if the responses are congruent and correct.(S7, T1) Comparing the two pictures, taking about the task, summing up differences.(S8, T2) We compared each other's pictures and tried to decide if we had understood each other.(S15, T2) On competing the task I did not use any strategy.(S 19, T1 T2)

A
talkative partner helped me a lot.(S1, T1) I was able to do the task thanks to my extensive knowledge of vocabulary.(S1, T2) What helped me the most was skillful asking of questions (e.g., asking for details), which enabled me to find out what I needed.(S3, T2)