Improving Learners’ English-Speaking Accuracy by Using Interrogative Constructions

ABSTRACT


Introduction
Many studies show that speaking accuracy plays an important role in Business English.Unlike writing accuracy (Le, 2022), speaking accuracy in Business English refers to the appropriation of lexical semantics, comprehensibility of sentence semantics, the equivalence between what speakers thought and expressed, the intelligibility of phonology, and acceptability of grammar (University of Cambridge ESOL Examination, 2012;Lambert & Kormos, 2014, p.3;Pica, T. et al, 1989, p.64;Kramsch, 1994).In a Business Environment, the accuracy of the speaker's information plays an important role in communication (Locke, 2009, p.391).Furthermore, how accurately you communicate at work is more vital than the speed (Schein, 2010, p. 57).Furthermore, grammatical accuracy is vital for learners to acquire English (Dawood, 2014, p.37).Although speaking accuracy is important, 70% of learners at Phu Em company have trouble expressing the interrogative sentences they would like to convey/ ask partners.They already have the ideas in mind, but do not know how to express them using English.Therefore, many solutions have been put forward to improve speaking accuracy such as corrective feedback (Chu, 2011), different types of tasks (Skehan & Foster, 2008), and planning (Ellis, 2009).However, they seem unsatisfactory in Business English because they neglect phrasal semantics or express precisely what they mean.
The term "Interrogative constructions", which are sub-branch of construction grammar (Goldberg, 2006), are generally understood as integrating syntactic, semantic, and contextual dimensions of interrogatives (Ginzburg, 2001, p.450).Interrogative constructions chosen among other types of construction grammar are because the learners in this context suffer from expressing interrogative sentences.One advantage of constructions is to reconstruct earlier stages of a language to better understand the mechanisms of language change, and the possible developmental paths of certain synchronic structures.(Ferraresi & Goldbach, 2008).Another advantage is that constructions provide learners with a better way of structuring information than traditional grammar (Littlemore, 2009, p.168).Hence, construction appears to be more suitable with Business English because grammar is not itself but also connects with phrasal semantics.
Using interrogative constructions in teaching and learning English speaking skills is proposed to improve speaking accuracy (Nakamura, 2008;Baicchi, 2016).However, it has not proven to be able to improve all aspects of speaking accuracy in Business English.As a result, this study is undertaken with the expectation of seeking the application of interrogative constructions in learning English speaking skills to improve learners' English-speaking accuracy.

Aims of the study
The study aims first to investigate whether using interrogative construction can improve learners' English-speaking accuracy in teaching and learning English speaking skills or not.The study also aims to collect the learners' relevant feedback on using interrogative construction.

The significance of the study
On a practical level, there are two benefits for both learners and teachers.First, this study can allow learners who lack a background in English to express their own interrogative sentences whenever they come up with an idea in their head.In other words, interrogative construction might be the foundation for those who would like to speak English at a higher level and become more fluent in English later.Second, the results of the study are to carry potential implications for teachers who want to use interrogative construction in teaching English speaking skills.
On a theoretical level, the research extends the way to improve speaking accuracy, especially the semantic aspect of speaking accuracy by interrogative construction.In fact, the result might be widely generalizable to other similar contexts.

The scope of the study
This research just focuses on "speaking accuracy of interrogative sentences" rather than "speaking accuracy in general" due to the fact that only interrogative construction is used in the treatment.Hence, even if the transcription on pretest and posttest is full, only interrogative sentences are taken to analyze and proceed to the next step.In addition, the interrogative construction in this research shows only some of the generalization of interrogative sentences, as there might be other special cases of interrogative sentences that the construction cannot cover, such as in non-conventionalized situations.

Speaking accuracy
English speaking accuracy is defined slightly differently according to its' dimensions: phrasal semantics (Kramsch, 1994) or vocabulary, grammar (Tawfik, 2022), and phonology (Brown, 2001, p.268).According to Riemer (2010, p.2), vocabulary and phrasal semantics are subelements of semantics.In short, speaking accuracy has three main elements: grammar, semantics, and phonology.In other words, theoretically, speaking accuracy is the ability to produce error-free sentences in speaking English in grammar, semantics, and phonology.Practically speaking, accuracy is evaluated within a construct called CAF (complexity, accuracy, and fluency).Speaking accuracy refers to "the ability to avoid performance error" (Skehan & Foster, 2008).
As learners in this study are adult workers who have worked with various people in various cultures and countries, the hard-to-get accuracy, such as speaking like native speakers, is not the priority.What the speaking accuracy here actually means to them is somehow related to the speaking message: whether they can express exactly the message they want in L2 according to what they thought and whether the listeners understand their message.For that reason, instead of referring to accuracy as "error-free" as above, it should be changed into something more practical and achievable to this study's target learners.In short, speaking accuracy in this study means the appropriation of lexical semantics (University of Cambridge ESOL Examination, 2012), comprehensibility of sentence semantics (Pica, T. et al (1989, p.64), the equivalence between what speakers thought and expressed (the expression criteria) (Kramsch, 1994), intelligibility of phonology (University of Cambridge ESOL Examination, 2012), acceptability of grammar (Lambert & Kormos (2014, p.3).
There are two ways to measure the speaking accuracy of phrasal semantics: the idea unit (Kroll, 1977, p.85) and C-unit (Pica et al., 1989, p.72).Regarding speaking accuracy of grammar and lexical semantics, there exist five main ways: the percentage of error-free clauses (Skehan & Foster, 2008, p.8), the proportion of error-free T units per T units (Lambert & Kormos (2014, p.3), the ratio of error-free AS units (Chu, 2011, p.456), percentage of correct verbs (Yuan & Ellis (2003, p.13), and proportion of error per 100 words (Polat & Kim (2014, p. 193).Concerning phonological features, many researchers use the software named Praat (Yan & Kim, 2018).In this study, it would be exhausting and difficult for the researcher to take into consideration measuring each element with each tool.Accordingly, only one tool should be chosen, which is the proportion of error-free AS units per AS units proposed by Foster et al. (2000, p.361) to measure elements in speaking skills.This tool is the most suitable because it mainly measures syntactic units, covers the weakness of the semantics unit, and covers the intonation unit (Foster et al., 2000, p.365).
Speaking skills in general and speaking accuracy in specific seem both include the speaker's meaning, the sentence meaning and the language.However, grammar only focuses mainly on the rules of language: on syntax alone (Carter & McCarthy, 2006, p.425), on contrastive analysis (Bùi, 2008), on the communicative grammar of English (Leech, 2008) -grammar in a specific context, or on the form-function relationship (Aarts, 2011).As mentioned in Section 1, adult learners in this study do not know how to link their meaning with grammar rules, form, and function.This means that speaking accuracy might require more than just grammar and rules, which refer to the relationship between form-meaning, or interrogative constructions.

The use of interrogative constructions
Interrogative constructions are a sub-branch of the construction grammar (form and meaning) (Traugott & Trousdale, 2013, p.8). Ginzburg (2001, p.450) defines interrogative constructions as "rigorously integrating syntactic, semantic, and contextual dimensions of interrogatives".Comparing the elements between interrogative constructions and the umbrella term "construction grammar", morphology and phonology are missing.These root from construction grammar include a wide range of constructions: word construction, phrase construction and sentence construction.In short, interrogative constructions is to integrate syntactic, semantic, and contextual dimensions of interrogatives (Ginzburg, 2001, p.450).
Practically, some researchers use construction grammar in learning speaking skills by memorizing words (Nakamura, 2008), remembering the type of construction grammar (Holme, 2010) and priming (Baicchi, 2016).These studies mostly emphasize the memorization of construction.However, it is obvious that speaking is not only knowledge (stored construction) but also a skill to use what the learners know (stored construction) to convey their ideas in various contexts (Bygate, 2016, p.50).This skill includes conceptualizer, formulator, and articulator (Bock & Levelt, 1994) and other terms such as concept, selection, and production (Bygate, 2016, p. 50).In short, the use of interrogative constructions in learning speaking skills goes through the first stage is to memorize the interrogative constructions, and the second stage is to sharpen the basic skill of using the stored interrogative constructions (Bygate, 2016).
The relationship between the use of constructions and speaking accuracy is shown through the following studies.Nakamura (2008) shows that learners who can memorize and verbalize the rules for appearance constructions get higher scores on grammatical accuracy.Baicchi (2016, p.189) uses priming constructions and confirms that explicit and frequent focus-on-form increases native-like accuracy.These above studies illustrate that using constructions might improve native-like and grammatical accuracy.It provides learners with the link between meaning and form (Traugott & Trousdale, 2013, p.8) as well.Thus, it seems that constructions are able to improve another aspect of speaking accuracy: the expression criteria.However, in the Vietnamese context, currently grammar is mainly taught as rules as mentioned above.Whether the use of interrogative constructions can improve speaking accuracy seems unknown.That is the reason why this study should be carried out to investigate whether the use of interrogative constructions might improve speaking accuracy or not.

Research Questions
In short, the present study aims to address the following questions: 1. Can using interrogative constructions improve learners' speaking accuracy in learning English speaking skills?
2. What is the learners' feedback on the use of interrogative constructions?

Pedagogical Setting & Participants
The study involved 30 participants out of the population of 60 sales employees from Phu Em Co., LTD.Phu Em company's employees were chosen because of their current need to learn English and the convenience of the researcher.
In an attempt to keep the reliability of the research, stratified sampling, which means that "each group contains subjects with similar characteristics" (Creswell, 2013), was the best match for this research.Stratified sampling was based on the learners' placement test results to allocate 15 participants into each group (CG and EG) based on learners' level.

Design of the Study
The research employed a mixed method, which combined "pretest-posttest-true-experiment" (Larsen-Freeman & Hlong, 2014, p.67) and qualitative data analysis.The reason for this choice was that the mixed method "provides a better understanding of the research problem and question than either method by itself" (Creswell, 2012, p.535).

Data collection & analysis
The data collection procedure was divided into three main stages: the collection of pretest and pre-questionnaire; the collection from posttest, post-questionnaire, and survey; and the collection from the interview.
The data analysis procedure included analyzing the quantitative, qualitative, and mixed-method data.First, the proportion of error-free AS units per AS units from the pretest -posttest and 5point Likert scale from the questionnaire were analyzed through SPSS 20 Independent Samples T-test and Pair Sample T-test.Second, the quantitative data from survey was analyzed through SPSS 20 descriptive statistics.Third, qualitative statistics from the survey interview was done through thematic analysis.Then mixed methods were conducted to compare and contrast quantitative with qualitative results.

Results on whether there exists effects on learners' interrogative speaking accuracy
As the difference in mean scores in Pretest between 2 groups might affect the conclusions on the actual effects of interrogative construction on speaking accuracy, the Independent Samples T-test was carried out to check the Pretest score between two groups (Table 1).
The sig (2-tailed) in Independent Samples T-test was 0.954>0.05(Table 1), which indicated that there was no difference in the mean scores in Pretest between both groups.In other words, the two groups CG and EG were somehow at the same level before the treatment.The results on whether there existed effects on learners' interrogative speaking accuracy were deeper investigated on its elements: semantics, grammar and phonology.
Regarding speaking accuracy of grammar, both groups experienced modest rise in mean scores (the change in mean score of CG=0.02, and in EG=0.01) (Table 2).This indicated that the CG learners improved more than EG learners regarding the acceptability of grammar.However, the treatment in both groups did not stress any significant difference on speaking accuracy of grammar (sig 2-tailed of CG=0.082, and of EG=0.504).The effect size of CG was 0.25 while the effect size in EG was 0.041.In short, the treatment did not bring about any important effect on acceptability of grammar of both groups; however, CG reflected noticeable improvement (25%) while EG constituted a modest improvement (4.1%).In the matter of the intelligibility of phonology, both groups did not bring about any significant effects after the treatment (sig 2-tailed in CG=0.72>0.05, in EG=0.55>0.05)(Table 3).The effect size of CG was 0.0119 while the effect size in EG was 0.0333.In short, the use of interrogative construction and the use of interrogatives produced no effects on speaking accuracy of phonology after 10 weeks of treatment, but slight improvement in CG (1.19%) and EG (3.33%).Regarding speaking accuracy of semantics, both groups reflected improvement in the mean score, which the mean score changed in CG is 0.03 and that in EG was 0.02 (Table 4).Moreover, CG reflected more considerable change in mean score than EG (0.03-0.02=0.01).However, there were no significant differences between pretest and posttest in both groups (sig 2 tailed in CG=0.52>0.05,sig 2 tailed in EG=0.65>0.05).The effect size of CG was 0.0386 while the effect size of EG was 0.0193.In other words, the treatment appeared not to bring about any substantial change on the speaking accuracy of semantics in both CG and EG groups.Nevertheless, the result yielded marginal improvement after the treatment, particularly the CG experienced the minimal effect of 3.86% while the EG produced negligible effect of 1.93%.Regarding the speaking accuracy in general between Pretest and Posttest, there was a modest rise in the mean scores in both groups, which suggested that the students in both groups did perform better in their Posttest (Table 5).The mean difference between its Pre-test and Posttest in CG was -0.07 while that of the EG was -0.079.However, the sig (2-tailed) in the CG (p=0.16 > 0.05) and the EG (p=0.08>0.05)conveyed that there were no significant differences in the mean scores of the Pretest and Posttest in both CG and EG group.The effect size of CG was 0.1719 and that of EG was 0.2789.In other words, the speaking accuracy of learners in CG improved 17.19% while that in EG improved 27.89% after the treatment; however, no considerable variations were found between these pairs.Regarding the general perception of learners (first 4 questions) towards the interrogative's expression criteria, changes in perception occurred in both groups (Table 6) through questionnaire.Specifically, CG reflected a minor change from disagree (M=2.37) to neutral (M=3.19).EG produced an overall change from "disagree" (M=2.44) to "agree" (M=3.73).This indicated that after the treatment, the CG learners might lack confidence while most EG learners thought that in general, they could express the questions they thought of.
Towards the other four questions of their substitution of expression (Table 6), CG remained neutral (M=3.19,M=2.75).Changes took place in EG from "agree" (M=3.5) to "disagree" (M=2.54).This showed that after the treatment, the CG learners still felt "neutral" about their use of these substitutions.On the contrary, EG learners thought they did not use these substitutions (body languages, interpreters, etc.) as frequently as before the treatment.Taken as a whole, these results offered evidence for the fact that the use of interrogative constructions improved the interrogative speaking accuracy in terms of interrogative expression criteria.

Results on EG learners' feedback on the use of interrogative constructions
According to the survey, when the subjects were questioned on the benefits of using interrogative construction, the majority commented that it resulted in a noticeable improvement in their grammar (M=3.92>3.4) and semantics (M==3.67>3.4)(Table 7).However, learners witnessed no improvements in their phonology (M=2.75).In other words, learners using interrogative construction felt unsure about its effects on the intelligibility of phonology.Interestingly, for the highest values of the benefits, grammar was found (M=3.92), which meant that learners mostly agreed that the use interrogative construction could improve their acceptability of grammar in context.This was a very contrasting result to the test score, which shows that learners in EG improved little after the treatment (4.1%) and also improved less than the CG group.
Table 7 Learners' feedback on the benefits and drawbacks of using interrogative construction Benefits Mean q.8a (semantics) 3.67 q.8b (grammar) 3.92 q.8c (phonology) 2.75 Drawbacks Mean q.9a (find pattern) 3.4167 q.9b (memorize pattern) 3.0000 q.9c (apply in context) 3.5833 The overall response to the difficulties of using interrogative constructions was quite negative (Table 7).Learners found difficulty in finding pattern of construction q.9a (M=3.42), and could not apply in speaking context q.9c (M=3.58).The majority of respondents showed a neutral opinion (M =3.00) towards the difficulty in memorizing the patterns of interrogative constructions.
However, their final opinion towards the use of interrogative constructions was nearly in between.Fifty-eight percent of those surveyed (58%) reported that they liked the use of interrogative constructions (Figure 1).
Figure 1 Responses regarding learners' opinions towards the use of interrogative constructions Pertaining to the benefits and drawbacks of using interrogative constructions, the interview showed similar codes as questionnaire but deeper understanding.
Furthermore, the open-ended questions of the questionnaire and interview also formulated 4 emerging codes in benefits of using interrogative constructions: self-reflect, thinking and speaking faster once the learner had a question in mind, more confidence in speaking, and guessing listeners' thoughts after learning interrogative constructions.The interview drew up 3 emerging codes towards the drawback of using interrogative constructions: difficulty in thinking and matching, memory issue and boredom.
In terms of the effect on learners, learner 29 and learner 30, who was EG immediate-level learners, reported that using interrogative constructions seemed to have no effects on them.They explained that they thought the language was more diverse while interrogative constructions were kind of fixed.

Discussion
The result (no effect size on grammar) supported Goldberg (2006), who indicated the overgeneralization of syntax from learners.However, it conflicted with the obvious conclusion of Nakamura (2008), which pointed to the fact that the use of construction grammar could improve grammatical accuracy.
The second major goal of this study was to collect learners' feedback on the use of interrogative constructions.Although the EG learners' feedback were that they achieved considerable Like 58% Dislike 42% Which of the below statements best express your opinion on the use of interrogative constructions?
improvements in their grammar and semantics, the actual results of their pretest and posttest did not reflect any important difference.This might be because learners appeared to be reasonably confident and because of the overgeneralization mentioned above.On the one hand, they might be able to recognize that their questions fitted in similar semantics or similar patterns of interrogative constructions, which led them to become entirely confident.On the other hand, because of that increasing confidence, they applied the structure of interrogative constructions in all situations despite the listeners and other elements.
Concerning the drawback of the use of interrogative constructions, the fact that learners encountered considerable difficulties in finding patterns of interrogative constructions was consistent with two general properties of language learners -conservative and attentive (Goldberg, 2006, p.64).The conservative learners would be better at learning examples.As a result, they somehow suffer from real trouble with finding construction patterns.

Conclusion
Although the study was able to answer the research questions, the findings might not be generalized to a larger context due to shortcomings below.First, the measuring speaking accuracy (AS unit) tool was not perfectly reliable.In order to minimize this weakness, the researcher used a questionnaire and interview to survey the learners' attitudes beside the statistical data.Besides, the questionnaire was carefully piloted to ensure its reliability.Second, another drawback of the research was the sample size.Although the learners were collected through stratified sampling, a weak form of true experimental design, which included 30 learners, was not valid enough to generalize the results for the whole population.
Third, the reliability of the research appeared to be influenced by one outside factor, which was Covid 19.It caused some lessons to be taught online.This did not only affect the learners' concentration on the lesson, but also influence the presence of learners in each section.Finally, as this was the first time the researcher handled teaching online without previous preparation, the quality of the lesson was also affected.
In short, the use of interrogative constructions does not bring about effects on grammatical elements, phonological elements, lexicons, and comprehension of speaking accuracy but only exerted effect on interrogatives expression on low level learners who were attentive.Moreover, the study reveals that half EG learners provide positive feedback on the use of interrogative constructions in learning English speaking skills.This indicates that interrogative constructions might not be suitable for all learners.The teachers should depend on learners' level and characteristics to decide whether using interrogative constructions is appropriate for them in learning English speaking skills.
This study focused only on the relationship between interrogative construction and interrogative speaking accuracy.In future research, another study should be conducted under different conditions, such as offline rather than online.Moreover, it would be better to carry out a study with larger population to cross-check the validity and reliability of the results.Another future direction might be the effects of using interrogative constructions on low-level adult learners' English interrogative expression to consider making generalization on the subject.

Table 1
Group Statistic of the CG and EG's Pre-test Scores (part 4).

Table 2
Comparison of the CG and EG's scores on the Pre-test and Post-test (part 1-acceptability of grammar)

Table 3
Comparison of the CG and EG's scores on the Pre-test and Post-test (part 2-intelligibility of phonology)

Table 4
Comparison of the CG and EG's scores on the Pre-test and Post-test (part 3-speaking accuracy of semantics)

Table 5
Comparison of the CG and EG's scores on the Pretest and Posttest (part 4-speaking accuracy in general)

Table 6
Comparison of the CG and EG's mean scores on the Pre-questionnaire and Post-questionnaire (part 2-in general)