Experiential versus attitudinal topic types and task performance in EFL monologues

With the aim of contributing to the existing ‎literature on the relationships between task ‎and topic facets, discourse features, topic ‎familiarity, and task performance in speaking, this study used EFL monologues to examine how two different sets of ‎topics―experiences/preferences versus opinions/attitudes―relate to task performance. The ‎task performance was measured using discourse features, including how language ‎elicited was complex, fluent, and lexically diverse. The study also explores how discourse ‎features themselves relate to one another across the two sets of topics. The data for the study ‎came from monologues performed by 63 adult EFL learners at the intermediate level of an ‎intensive English program in Saudi Arabia. The learners produced the monologues in response ‎to two summative tests (i.e., Test 1: experiences & preferences and Test 2: opinions & ‎attitudes). Using parametric statistical analyses (incl., the paired samples T-test and the ‎Pearson correlation), it was found that while experiences and preferences evoked more fluent ‎language than did opinions and attitudes, the latter elicited more complex and lexically ‎diverse language. Also, a significant, positive correlation existed between fluency and complexity for experiences and preferences, whereas lexical diversity was significantly positively correlated with complexity for opinions and attitudes. The study report concludes ‎with practical implications for enhancing task performance of monologues in the areas of ‎complexity, fluency, and lexical diversity.‎


INTRODUCTION
The importance of speaking ability in English for EFL learners cannot be stressed more. Candidates who are highly proficient in English speaking can do well on high-stakes tests, have good chances of finding jobs, and study in English-speaking countries. There are a number of variables that shape oral performance at the level of both overall oral ability and its componential skills. Such variables include task design or format, topics or prompts, and scoring or rating criteria (Assiri, 2017). Task facets can influence both how much a learner can produce orally and how well he can perform. According to Skehan (2014), numerous practical implications offered by studies that addressed aspects of oral tasks have proven useful to classroom practice.
The extent of background knowledge relating to a topic determines its level of familiarity. Not having an adequate level of prior knowledge about a given topic can make a learner's task of talking about it perplexing (Robinson, 2011). The fact that a certain topic is familiar to a learner means that the learner can retrieve its related information quickly and easily. As such, the learner can be said to be conceptually prepared to perform the oral task at hand (Skehan, 2016). Even on high-stakes tests, test takers often report that the topics included in the speaking parts of those tests are not familiar to them (Smith, 2009). This, in turn, makes test takers feel anxious and unable to perform well.
Discourse features comprising complexity, fluency, lexical diversity, and accuracy represent the main foci in the most recent research in task-based language teaching (TBLT). They reveal significant differences due to effects of test formats better than do task ratings or scores (Gan, 2013). Gan maintains that the interpretations that we base on test scores can be made more valid if the discourse features associated with task performance are attended to. Our analysis of discourse produced in oral tasks can lose invaluable information if it lacks adequate consideration of the features of such discourse (Leaper & Riazi, 2014). Bui and Skehan (2018) suggest that the use of discourse features across a number of studies can make research findings comparable and of practical value. Nevertheless, how various task and topic types differ in their performance remains worth more exploration (Qiu, 2019). Other researchers (e.g., Bui & Huang, 2018;Qiu & Lo, 2017;Qiu, 2019) have called for research which aims to investigate such discourse features as complexity, fluency, lexical diversity , and accuracy in the light of task performance and in relation to topic familiarity. As a recent trend, researchers have shifted their attention from test scores to discourse features associated with oral production.
A task represents a unit of L2 lesson planning, teaching, and learning (Long, 2015). When learners perform L2 oral tasks, they go through stages that are best captured by Levelt's speaking model (Levelt, 1989). As Skehan (2018) mentions, the Leveltian model helps analyze the cognitive processes involved in L2 oral production while considering any factors pertinent to task design. In view of this framework, an L2 learner retrieves the background knowledge related to what he wants to say, decides how much information is needed and in what order this information will flow, activates the syntactic structures that match the semantic load of the information, and actuates the morpho-phonological codes in preparation for the physical production of an utterance. Pang and Skehan (2014) distinguish between learners who are at high-versus low-proficiency levels in terms of going through these stages such that while low-proficiency learners work more with the grammatical and morpho-phonological coding, highproficiency learners focus more on communicating their thoughts meaningfully. In order for learners to automatize these steps, tasks should be made increasingly complex using effective sequencing procedures (Lambert & Robinson, 2014). When learners are able to transfer their experience of dealing with a task as an instance to other tasks with similar structure, this creates a sense of task familiarity (Bui, 2014).
So far, two competing views seeking to explain the relationships between task facets and discourse features have been put forward in the TBLT domain. The first is Skehan's (2009) limited processing capacity (or tradeoff) hypothesis, which suggests that because L2 learners typically exhibit limited processing capacity, they have to prioritize any of the performance elements of complexity, fluency, lexical diversity, and accuracy and that occurs at the expense of an(other) performance element(s). The other is Robinson's (2011) cognition hypothesis, proposing that a performance element like complexity can spur all or any of the other elements including fluency, accuracy, and lexical diversity. Nonetheless, each view seems to speak to a certain task with distinct design factors that influence how the performance areas relate to one another. And so, the two views may not be necessarily competing after all, but rather complementary.
The performance elements, including complexity, fluency, lexical diversity, and accuracy are referred to in the literature as discourse features or aspects. Complexity denotes the degree to which a learner makes diverse and structured use of her L2 (Ellis, 2012). Fluency is the quality of producing rapid and fluid oral language (Segalowitz, 2016). Lexical diversity is defined as the extent to which the words in a given discourse are wide-ranging (McCarthy & Jarvis, 2010). Accuracy stands for the ability to produce well-formed or error-free language (Ellis, 2003). These discourse features tend to be perceived as independent entities, whereas in fact they work together in the manner they relate to task facets, as research findings (incl., Iwashita, Brown, McNamara & O'Hagan, 2008) point out. As is the case with many phenomena in L2 learning, the use of discourse features is moderated by individual differences that do not lend themselves to intuitive understanding (Lambert & Robinson, 2014).
Several studies have explored how task and topic types affect oral performance with regard to complexity, accuracy, and fluency. In this respect, Gan (2013) compared group interaction and individual presentation using a sample of 30 learners of ESL. The researcher found that the group discussions performed less than the individual presentations along the three dimensions of complexity, fluency, and accuracy. Gan explained this finding, suggesting that while delivering their presentations, learners did not experience as much pressure as they did in the group discussions, which made them perform much better. Similarly, Ahmadi and Sadeghi (2016) used monologue, interview, and group task formats with 23 EFL learners. They found that the language used in the interview and group tasks was less complex than that used in monologues. In another study, Leaper and Riazi (2014) had 141 EFL learners participate in group discussions using four topics: mobile, outdoors, singles, and family. The researchers observed more fluent interactions associated with the mobile and outdoors prompts when compared to the singles and family ones, and more complex exchanges with the latter set of prompts than with the former. These findings point to the need for more research involving other task and topic types so that we can broaden our understanding of how such factors shape oral performance At the heart of the research on the relationship between task and topic factors and oral performance lies the issue of topic familiarity. Topic familiarity is conceived as a form of implicit planning, which when exists, entails learners' readiness for the assigned oral task (Bui & Huang, 2018). As such, it manifests in a Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN: 2301-9468 graded, not divided mode (a topic may be partially familiar, but not only familiar or not familiar at all) (Qiu, 2019). Topic familiarity relates to the psychological states of confidence and ease upon which success in task performance is highly reliant (Teng, 2016). Bui (2014) considers topic familiarity as one dimension of task-internal readiness, which can make tasks assimilate to real-life experiences. The degree to which a task is structured can have a positive effect on task performance as assessed by discourse features (Tavakoli & Skehan, 2005). As opposed to unstructured tasks, structured tasks possess a more logical order of events or steps and are less cognitively demanding.
Generally, previous research has concluded that topic familiarity has positive effects on oral performance. Bei (2010) investigated the relationship between topic familiarity and oral performance among 80 learners of EFL and found that topic familiarity enhanced fluency, accuracy, and lexical diversity. Qiu and Lo (2017) examined if each of topic familiarity and task repetition had any effects on the manner in which 60 EFL learners carried out four narrative tasks-two of the tasks were on familiar topics and the two others were on unfamiliar topics. The researchers found the learners to be more engaged in the familiar tasks in comparison to the unfamiliar ones. The tasks with the familiar topics were shown to be interesting, elaborative, inviting for sharing while the ones with the unfamiliar topics were frustrating, worrying, and challenging. In the same vein, Qiu (2019) duplicated the Qiu and Lo's (2017) yduts with 60 EFL learners using four monologues. She noticed a strong link between expressing familiar topics and using complex language. More specifically, the familiar topics facilitated both retrieval of relevant information and organization of talks, whereas the unfamiliar topics conflicted with complex and fluent expression. Bui and Huang (2018) studied effects of pre-task planning and content familiarity on fluency with a group of 58 learners of EFL. The researchers found that compared with the unfamiliar topics, the familiar topics encouraged the learners to produce longer stretches of discourse. The familiar topics were also more associated with improvements in many aspects of fluency, including speech rate, pauses, and repetitions.
The findings from the previous studies (Ahmadi & Sadeghi, 2016;Gan, 2013;Qiu, 2019) suggest that monologues can outperform other task types (incl., dyadic and group tasks) in the way they affect task performance positively. In fact, scholars (e.g., Ockey, Koyama, & Setoguchi, 2013) note other merits of monologues. First, monologues are not subject to effects from an interlocutor or partners in the task. And, learners assigned monologic tasks are not exposed to the kind of communicative pressure that could influence their performance in a negative manner. Besides, monologues are characterized by practical administration in that the whole task can be audiorecorded, and so, it does not demand the presence of an interlocutor.
Another line of research has looked at the relationships between such factors as time-pressured tasks, topic preferences, engagement, and task performance. Thai and Boers (2016) asked 20 EFL learners to deliver monologues about movies they enjoyed the most. This was done under two conditions, one being shrinking-time pressure and the other constant-time pressure. The researchers found that, in the shrinking-time condition, fluency was augmented more than was each of complexity and accuracy. In fact, complexity and accuracy gained more in the constanttime condition. Phung (2017) focused on how task preferences related to engagement among 21 ESL learners. The researcher observed that the learners' affective reactions to their preferred tasks were better than those linked to the tasks they disliked. He also noted that the learners' engagement in the tasks was considerably determined by the level of preference they assigned to these tasks.
In the light of the discussion above, this study aims to add to the existing literature on the relationships between task and topic facets, discourse features, topic familiarity, and task performance. In so doing, the study attempts to address three issues. First, it focuses on how topic types, including experiences and preferences versus opinions and attitudes, differ in regard to complexity, fluency, and lexical diversity. Then, contrary to most of the previous studies, it takes into account lexical diversity as a discourse feature in exploring how the two sets of topics relate to oral performance. Last, it looks into how discourse features relate to one another across the two types of topics.
Accordingly, the current study addresses the following research hypotheses (1-3) and questions (4&5): 1. Opinions and attitudes elicit more complex language than do experiences and preferences. 2. Experiences and preferences elicit more fluent language than do opinions and attitudes. 3. Experiences and preferences elicit more lexically diverse language than do opinions and attitudes. 4. How do discourse features for the experiences and preferences set of topics relate to one another? 5. How do discourse features for the opinions and attitudes set of topics relate to one another?

METHOD
Here is a detailed description of the method followed in this research.

Setting
The data were collected at an intensive program of the English language in Saudi Arabia. The program offers English skills to its students in order for them to be able to enroll in degree programs. It has four levels spanning eight months of language study with the focus on five Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN: 2301-9468 language skills: listening, oral, reading, writing, and grammar.

Participants
The participants in the study were 63 EFL learners, 48 males and 15 females, aged from 20 to 24, distributed among four oral classes. They were all at the intermediate (or third) level of the program. The participants can be said to represent a relatively homogeneous group concerning their mother tongue (Arabic), exposure to English, and cultural background. They were engaged in this research on the basis of convenience sampling.

Data Collection
During their study, students took two main tests (Test 1 and Test 2), Test 1 in the middle of the course and Test 2 at the end. As Table 1 shows, Test 1 included three topics classified as experiences and preferences and Test 2 had 3 topics grouped as opinions and attitudes. Both sets of topics were covered in the class materials. Before each test, students were provided with a list of topics, including the ones to be presented in the test, so that they could prepare for the test well. They were also instructed about how the test would proceed. During the test, each student had three minutes to talk about each one of the three topics. All the monologues delivered by students were audiotaped.

Data Analysis
Monologic data were transcribed and analyzed as ASunits following Foster, Tonkyn, and Wigglesworth's (2000) guidelines. The sample size (63) permitted use of parametric statistical tests (incl., the paired samples Ttest and the Pearson correlation). Table 2 presents how the three discourse features in the study were operationalized. They were measured for each monologue as follows: 1. Complexity: The number of words per AS-unit was calculated. 2. Fluency: The lengths of pauses were estimated using the Audacity software. 3. Lexical diversity: The vocd-D, a measure of vocabulary richness, was obtained using the VOCD software. 1) Talk about transportation in the Kingdom. 2) Discuss your future job.
2) Describe a problem affecting the world today. 3) Describe your best friend.
3) Discuss the risks of using IT.  Table 3 displays the means and standard deviations of all the measures across the two sets of topics. To find the answers to the three hypotheses (1-3), paired samples T-tests were run. The assumptions of these statistical tests were satisfied. Both Test 1 and Test 2 were administered under the same conditions. The mean differences for all the measures were normally distributed.

RESULTS AND DISCUSSION
With respect to the first hypothesis "opinions and attitudes elicit more complex language than do experiences and preferences", the words/AS-units measure of complexity pointed to a significant difference that existed between the opinions and attitudes (M=12.699, SD=2.706) and experiences and preferences (M=10.518, SD=2.815) sets of topics; t(62)= 4.186, p=.000. Therefore, opinions and attitudes elicited more complex language than did experiences and preferences. The result of testing the first hypothesis suggests that monologues on opinions and attitudes were more complex than were those on experiences and preferences.
It might be because learners were more engaged in critical thinking when expressing their opinions and attitudes as opposed to experiences and Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN: 2301-9468 preferences. Therefore, with the latter, all that learners had to do was to retrieve the background information that allowed them to reflect on their experiences or describe their preferences. Conversely, when they expressed their opinions and attitudes, they had to make evaluations and take a stand about the topics requiring their views. Another explanation may have to do with topic difficulty. Since opinions and attitudes are presumably more difficult and challenging to learners than were experiences and preferences, they demanded higher levels of complexity. This finding accords with Robinson's (2007) model which suggests that tasks whose content is demanding elicit highly complex language. Although opinions and attitudes were supposedly less familiar to learners than were experiences and preferences, learners managed to perform well while delivering them. This can be attributed to the fact that learners made use of planning in the process of getting ready for their monologues. As Bui and Huang (2018) suggest, planning helps learners become more adjusted to the demands of unfamiliar and challenging tasks and topics.
As for the second hypothesis "experiences and preferences elicit more fluent language than do opinions and attitudes", lengths of pauses indicated a significant difference between the opinions and attitudes (M=94340.286, SD=11310.528) and experiences and preferences (M=87869.540, SD=16021.135) sets of topics; t(62)= 2.479, p=.016. Considering the fact that opinions and attitudes had longer pauses compared to experiences and preferences, we can say that the latter set of topics called for more fluent language that did the former.
The second hypothesis is confirmed, which implies that learners generally handled experiences and preferences much more smoothly than they did with opinions and attitudes. Because opinions and attitudes were presumably less familiar to leaners than were experiences and preferences, they may have used their pauses while performing the former set of topics to do extensive online planning. As was the case with complexity being more typical of opinions and attitudes due to their difficulty, the easiness of experiences and preferences played a role in linking them to fluent production. This gives evidence to the hypothesis that learners were more familiar with experiences and preferences than they were with opinions and attitudes . The more learners are familiar with topics and contents of the oral tasks they will perform, the more they will feel engaged and become productive during these tasks (Qiu & Lo, 2017) . This seems to apply to the manner in which learners in this study dealt with their experiences and preferences. Also, this points out the role of topic familiarity in maximizing language processing as it facilitates drawing on relevant cognitive resources, as Qiu (2019) notes.In regard to the third hypothesis "experiences and preferences elicit more lexically diverse language than do opinions and attitudes", the measure of vocabulary richness (vocd-D) showed a significant difference between the opinions and attitudes (M=40.000, SD=9.978) and experiences and preferences (M=36.400, SD=9.458) sets of topics; t(62)= 2.194, p=.032. In other words, opinions and experiences invited more lexically diverse language than did experiences and preferences.
The result of testing the third hypothesis suggests that when learners reacted to the topics drawing on their opinions and attitudes, they made more use of a variety of vocabularies than was the case with experiences and preferences. In fact, with the latter set of topics, they generally used more words (i.e., on average 146 versus 128), but these words were mostly not of unique types. It can be assumed that because opinions and attitudes were more complex than were experiences and preferences, learners were led to use more diverse vocabulary with the former set of topics than with the latter. To reiterate, complexity in the answer to the first hypothesis was explained in terms of topic difficulty. Therefore, due to the fact that opinions and attitudes were more difficult to learners than were experiences and preferences, they were linked with more diverse vocabulary use. Similarly, Robinson (2007) found that complex narrative tasks ‫‬ called for more lexically diverse language from the learners in his study. Although opinions and ‫‬ attitudes were apparently more difficult than experiences and preferences, they did not impede ‫‬ lexical diversity in this study as did the difficult tasks used in ‫‬ Préfontaine and Kormos (2015).
Pearson correlations were used to answer the research questions (numbered 4 and 5). The results of the answer to the research question "how do discourse features for the experiences and preferences set of topics relate to one another?" are shown in Table 4. Only one significant positive correlation existed between complexity and fluency as a result of presenting the experiences and preferences set of topics, r(63) = 0.257, p=.042. The answer to the research question about the relationships between discourse features associated with the topics on experiences and preferences indicate that only complexity and fluency were positively related. Therefore, there is a fixed relationship between the two discourse features for this set of topics such that as complexity went up, fluency went up, too. Learners made longer pauses as they were trying to produce complex speech. Fluency was clearly more sensitive to changes in complexity than was lexical diversity. The positive association between complexity and fluency of the discourse produced in response to familiar topics was also observed by Bui (2014). Leaper and Riazi ‫4102(‬ ‫‬ ( ‫‬ also noticed that their respondents used fluent and complex ‫‬ language when talking about their future plans. This result aligns with the answer to the second hypothesis in that experiences and preferences were produced fluently. Lexical diversity did not show to be a high correlate with complexity or fluency. In fact, a negative correlation was found between complexity and lexical diversity, although non-significant . Also, the correlation between fluency and lexical diversity was non-significant. Table 5 displays the results of the answer to the research question "how do discourse features for the opinions and attitudes set of topics relate to one another?" The only significant positive correlation existed between complexity and lexical diversity for the opinions and attitudes set of topics, r(63) = 0.430, p=.000.
The answer to the research question about the relationships between discourse features linked with the topics on opinions and attitudes suggest that only complexity and lexical diversity were positively related. Therefore, with this set of topics, as complexity went up, lexical diversity went up, too. As learners were trying to produce complex speech, they made more use of a variety of lexical items. In this respect, lexical diversity was obviously more sensitive to changes in complexity than was fluency. This result is compatible with the answer to hypothesis three in that opinions and attitudes resulted in language that was lexically diverse. Fluency did not appear to be a high correlate with complexity or lexical diversity. In fact, a negative correlation was observed between complexity and fluency, although non-significant. Also, the correlation between lexical diversity and fluency was non-significant. As opposed to Qui (2019) who found that lack of topic familiarity interfered with lexical diversity, this study found that opinions and attitudes which were presumably low on the familiarity scale compared to experiences and preferences prompted more lexical diversity. Gan (2012) also found a positive association between complexity and lexical diversity in the language used by his participants while giving presentations on complex topics. Note. **: significant at .01 (2-tailed)

CONCLUSIONS
The findings of this study suggest that the two sets of topics (i.e., experiences and preferences versus opinions and attitudes) differ in eliciting oral language in monologic tasks. To be specific, while opinions and attitudes call for complex and lexically diverse language, experiences and preferences produce fluent speech. This can be due to the differences between the two sets of topics in their cognitive demands, with experiences and preferences being descriptive in nature whereas opinions and attitudes critical. Also, opinions and attitudes are conceptually more difficult, and less familiar, than are experiences and preferences. Another consideration is the fact that since opinions and attitudes demand higher levels of complexity than do experiences and preferences, they also induce the use of more lexically diverse language. The relationships between discourse features are variable and dependent on task facets, including topic type. With topics that are simple and familiar, learners are tempted to produce language that is both complex and fluent. Such a level of complexity is not associated with a high level of lexical diversity. However, learners tend to produce complex and lexically diverse language when dealing with difficult and unfamiliar topics. High levels of difficulty and unfamiliarity of topics limit fluent expression. The fact that experiences and preferences have high fluency, but low lexical diversity, in relation to complexity gives evidence in this study for Skehan's (2009) tradeoff hypothesis . Similarly, for opinions and attitudes to have high lexical diversity, but low fluency, in relation to complexity gives another piece of evidence in support of the trade-off hypothesis. At the same time, Robinson's (2011) cognition hypothesis partially applies in that with the first set of topics complexity went hand in hand with fluency, and with the second set of topics in that complexity and lexical diversity worked together. In both cases, complexity did not work with more than one discourse feature. In brief, topic types can determine the nature of the trade-off that can occur to discourse features in monologic tasks and Copyright © 2018, IJAL, e-ISSN: 2502-6747, p-ISSN: 2301-9468 which discourse features work in tandem. These findings have major implications that will be discussed below.
Learners ought to be familiarized with the topics and contents of the monologues they expect to deliver by brining to their attention some aspects of the background knowledge they will need. This will ensure more unified performance by controlling for any potential discrepancies among learners in this regard.
i , pmaxe r F e m,e ,y m, my t d t eFe , a e e cu y e d F, me ,Fd mgdFeFdF yr d s gme n x, eFt thFd pmaxe y em,F uymgdFeFdF yd mdx xe typically enjoy my h ee my d F, , e emed e gmnuem,F y met , udFe y. Tasks and topics can be arranged so that they flow from simple to challenging ones (Bui & Teng, 2018). This can be applied to both teaching and assessment, for instance, topics that relate to learners' preferences should precede those topics that require their evaluations and critical thinking. Unfamiliarity with topics can have such detrimental effects on performance of monologues that should be avoided at any cost. Not ensuring adequate level of topic familiarity to all students can result in our teaching and assessment practices being biased and futile (O'Sullivan & Green, 2011). Care should be taken not to select topics that are intrusive or private in nature as this may influence students' task performances negatively (Leaper & Riazi, 2014). Examples of intrusive topics include asking learners to talk about their marital lives or familial circumstances. Generally, as Préfontaine and Kormos (2015) suggest, when it comes to task design and topic selection, teachers are advised to make careful decisions and choices that would ultimately help leaners develop their oral ability.
Topic familiarity can be ascertained in other ways. Teachers ought to discern the levels of background knowledge among their students and use this understanding as a criterion for topic selection (Bui, 2014). s Fy gme n t e ns yu,e sFec ydut edy' Ft my met px ,F eg y Fe , emdF e d m em,F s x d edFme d xFgy. s mg ,y can also have their students repeat the oral tasks required of them, which can boost their familiarity with the contents of these tasks (Qiu, 2019). dut edyamsd,sd e ha , mn udd d xFgy myyFce td d a ed nmyFy d F, F,ydx , ,ameg ns FetFec a , , e emed Fe ,amdF e. Teachers may also have their students use a multi-question format when talking about a certain topic so as to minimize the effect of the topic being unfamiliar or challenging to students (Khabbazbashi, 2017 (Lambert, Philp, & Nakamura, 2017). Learners could be encouraged to propose or select topics they find familiar and engaging (Phung, 2017), for instance. Topics unfamiliar to learners are not likely to be relevant to their daily life, and so they cannot be beneficial to their overall learning experience. One major limitation in this study is that it targeted EFL learners at the intermediate level. This limits the generalizability of its findings to learners at different proficiency levels. Therefore, more research is recommended with learners at various proficiency levels, including preparatory, elementary, and advanced. This will make it possible to compare research findings across proficiency levels. It will also build our understanding of the intricate relationships between topic types, discourse features, and oral production as far as TBLT is concerned. Another limitation is that this study did not make use of self-report methods. Use of such methods would have informed the study design in regard to the extent to which learners found the topics to be familiar or challenging. Also, a combination of stimulated recalls, self-observations, and retrospective interviews could have provided full accounts of learners' reactions and response behaviors in relation to the topics assigned to them (see Assiri, 2011Assiri, , 2016 for an example of how this combination can be useful).