147 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... Birtan Baytar UDK [378-057.875:81]:81'246 Department of Foreign Languages DOI: 10.4312/vestnik.14.147-176 Kastamonu University Izvirni znanstveni članek Turkey bbaytar@kastamonu.edu.tr İsmail Çakır English Translation and Interpretation Ankara Yıldırım Beyazıt University Turkey icakir@ybu.edu.tr A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL TERTIARY LEVEL STUDENTS ON THE RELATIONSHIP BETWEEN ANALYTIC LANGUAGE KNOWLEDGE AND METALINGUISTIC AWARENESS* 1 INTRODUCTION 1 The relationship between metalinguistic awareness and previous language knowledge has been a topic of research. Nevertheless, the effects of language knowledge based on explicit formal language knowledge have not been studied in detail. Moreover, research on bilinguals has had some controversial outcomes, especially that carried out in an immigrant context. Specifically, some studies found negative results for bilinguals, and it was thought that being bilingual had detrimental consequences for the cognitive development of children, and thus was not favoured among experts in this field (Hakuta & Garcia, 1989). However, these studies did not reflect the reality well, as the instruments used and variables such as the socio-economic backgrounds of the subjects were not taken into consideration (Barac & Bialystok, 2011; Hakuta, 1986; Hakuta & Garcia, 1989; Hakuta & Suben, 1985). Analytic language knowledge and metalinguistic awareness have also been dis- cussed with regard to whether they have any effects on each other, although this relation- ship has not been studied in depth among Turkish tertiary level students. Monolinguals and bilinguals have their advantages and disadvantages in different tasks. However, how they perform in tasks requiring metalinguistic awareness remains unanswered, which thus forms the main issue examined in the present study. 1 This manuscript is based on the first author's master thesis, which was written under the supervision of the second author. Vestnik_za_tuje_jezike_2022_FINAL.indd 147 Vestnik_za_tuje_jezike_2022_FINAL.indd 147 24. 01. 2023 09:18:49 24. 01. 2023 09:18:49 148 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES In line with the situation outlined above, the relationship between the level of metalin - guistic awareness of the students and analytic language knowledge was measured using a test adapted from Ter Kuile et al. (2011), in which there were 15 questions written in Indonesian (i.e., a language that was not known by any of the participants in the study). The students were from the departments of History and Philosophy, Turkish Language and Literature, and Eng- lish Language Teaching and English Language and Literature. Demographic data was first gathered, followed by the students taking the metalinguistic awareness test, with the whole process implemented by the researcher and taking around 15 minutes in total. The gender, age, exact length of language learning, the period of language learning, attending a prepara - tory class, and the proficiency level of the participants were asked, as well as any languages known by the students other than Turkish and English. The answers to the demographic ques - tionnaire were also analysed to investigate if there were any relationships among the variables, and if so which variable had an effect on the others. Overall, this study aims to find answers to the following questions; Is there a relationship between metalinguistic awareness and analytic language knowledge? What is this relationship for students from the departments of His - tory, Philosophy, Turkish Language and Literature, English Language Teaching and English Language and Literature? Do gender, age, age of starting language learning, previous years of language learning, preparatory class, proficiency level and knowing any languages other than Turkish or English have an effect on metalinguistic awareness levels of the participants? 2 LITERATURE REVIEW Multilingualism is becoming more common due to globalization, and thus the next section explains the concepts of monolingualism, bilingualism, multilingualism, sequential bilin - gualism, metalinguistic knowledge, metalinguistic ability, metalinguistic awareness and the threshold level hypothesis, which will be frequently encountered throughout the current work. 2.1 Monolingualism, Bilingualism, Multilingualism In general terms, being monolingual means knowing only one language, but according to Ro - maine (2013) and Paradis (2004), today most people speak more than one language, and thus they are bilingual, and speak two languages, or multilingual, and speak three or more. There - fore, bilinguals and multilinguals can be considered as the majority, rather than the minority. 2.2 Sequential Bilingualism Sequential bilingualism, as implied by the study’s title, is a central concept in the current work. Sequential bilingualism means that there is an order in the acquisition of languages. Vestnik_za_tuje_jezike_2022_FINAL.indd 148 Vestnik_za_tuje_jezike_2022_FINAL.indd 148 24. 01. 2023 09:18:49 24. 01. 2023 09:18:49 149 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... After one of the languages is learnt completely, learning of a second language starts. The first language here is basically the mother tongue, meaning that it is the language spoken at home and acquired as a child in a natural way, whereas the second language may be acquired as a result of a formal education at school or in a natural way again, as a result of migration, adoption etc. (Berken et al., 2017). 2.3 Metalinguistic Knowledge The understanding of grammar, form, and structure is referred to as metalinguistic knowl- edge. Bialystok (2003) states that metalinguistic knowledge should include information about canonical word order and productive morphological patterns. Understanding these abstract concepts is different to understanding a specific language, and metalinguistic fea - tures have a great influence on language (Kopečková, 2018; Mertz & Yovel, 2009). Falk, Lindqvist and Bardel (2013) carried out a study in order to observe the effects of metalin - guistic knowledge of the mother tongue on third language oral production, and found a pos - itive correlation between metalinguistic abilities and various tests applied during the study. 2.4 Metalinguistic Ability The terms metalinguistic knowledge and ability are closely related to each other, and in fact refer to the same issue from different perspectives. Bialystok (2003) describes metalinguistic ability as a reward that comes to those who have earned it, by being, for example, multilingual, more educated and intelligent. Moreover, metalinguistic skills are linked to cognitive development, with a positive relationship between them. Therefore, in order to deal with metalinguistic problems and solve tasks requiring metalinguistic abili- ties, learners should aim to improve their cognitive skills (Bialystok & Bouchard, 1985). 2.5 Metalinguistic Awareness Metalinguistic awareness means specific attention that is focused on the language being used at the time of speaking (Bialystok, 2003). Adult learners have an advantage in terms of metalinguistic awareness when compared to younger learners. This is because adult learners are cognitively developed and can use advanced analytic problem-solving abili- ties that broadly pave the way for metalinguistic knowledge (Roehr, 2007). Additionally, there is a strong connection between metalanguage and metalinguistic knowledge. While metalanguage refers to all the terms used to explain language, metalin- guistic knowledge refers to explicit understanding about language (Hu, 2011), and these two terms are interdependent. Vestnik_za_tuje_jezike_2022_FINAL.indd 149 Vestnik_za_tuje_jezike_2022_FINAL.indd 149 24. 01. 2023 09:18:49 24. 01. 2023 09:18:49 150 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES 2.6 The Threshold Level Hypothesis Cummins (1976) examines why studies concerning the effects of bilingualism on cog- nitive and linguistic skills revealed different results, and proposes “the threshold level hypothesis”, saying that bilinguals should achieve a certain level of proficiency in both languages to get the advantages of being bilingual, otherwise knowing two languages might not affect – or may even adversely affect – the cognitive development and linguis- tic skills of the learners. Moreover, this work explains that learners should achieve this threshold level not just to enjoy the advantages of knowing at least two languages and thus find additional language acquisition easier, but also to avoid any possible cognitive problems, as there might be negative effects of bilingualism as well as positive ones un- less the threshold was attained (Cummins, 1976). The threshold level can tell us about the cognitive functioning and language abilities of bilinguals, and vice versa (Cummins, 1979; Hakuta, 1987). Cenoz (2003) presents the importance of the place of languages in bilingualism, and states that when L1 is not replaced by the second language, positive effects can be expected. Gonzalves (2020) also points out that L1 literacy affects metalin- guistic awareness in the second language. Therefore, Cenoz (2003) favours the threshold level hypothesis and its subdivisions, namely the upper and lower thresholds, with the former referring to balanced bilinguals, with positive effects expected for the cognitive and linguistic development of learners, whereas the latter refers to a low level of profi- ciency and no or even negative effects on the learners. 3 METHODOLOGY Two different instruments were used in order to gather the data, a metalinguistic aware- ness test and demographic questionnaire. Descriptive statistics, one-way ANOV A and Kruskal-Wallis tests were used to analyse the data. The participants were all tertiary level students from three different Turkish state universities. 3.1 Research Design Quantitative research was adopted in this study in order to gain the information needed to better understand the relationships among various factors and the participants’ metalin- guistic awareness (Brown & Rodgers, 2003; Mackey & Gass, 2005). The descriptive sta- tistics of the items, one-way ANOV A and Kruskal-Wallis test results were all evaluated using IBM SPSS 22. The reliability of the study, item discrimination and item difficulty were analysed using the necessary formulas in Microsoft Excel (Atılgan, Kan & Doğan, 2007). The study was carried out with 210 tertiary level students studying in three differ- ent state universities in Turkey. Vestnik_za_tuje_jezike_2022_FINAL.indd 150 Vestnik_za_tuje_jezike_2022_FINAL.indd 150 24. 01. 2023 09:18:49 24. 01. 2023 09:18:49 151 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... 3.2 Data Collection Tools 3.2.1 Metalinguistic Awareness Test An Indonesian Language Test (ILT) adapted from Ter Kuile et al. (2011) was used as a tool to evaluate the metalinguistic levels of the students, in which there was a short read- ing text (183 words) in Indonesian (see appendix). Indonesian was chosen because none of the participants knew this language, and the test’s questions are simple and based on analytical skills. The reason for choosing an unknown language was to eliminate the risk of answering the questions in the metalinguistic awareness test without using any meta- linguistic skills. The students were asked to answer 15 questions about the story told in the text, the questions were about meaning, grammar, and vocabulary. The given words included key terms for the text. The questions could all be answered with the help of the students’ previous language learning knowledge and experience (if possible). There are some difficult questions as well as relatively easy ones, and the difficulty level of the questions and their level of discrimination are analysed in the item analysis section. 3.2.2 Demographic Questionnaire Along with a consent form stating that each subject engaged voluntarily in the study, the students received a questionnaire requesting some background information. 3.3 Participants and Setting A total of 210 tertiary level students from three public universities and five different depart - ments participated in this study voluntarily. While the participants were selected randomly in their specific majors, the majors were selected on purpose to draw a parallel between the groups and participants. As two participants were found to be outliers in the analysis sec - tion, they were excluded from the study. Therefore, 208 students participated in the actual study, 162 of whom were female and 46 were male. Moreover, the students were from the departments of History, Philosophy, Turkish Language and Literature, English Lan- guage Teaching and English Language and Literature. Although the participants are from five different departments, they are divided into three subgroups ( History and Philosophy, non-analytical language study group; Turkish Language and Literature, monolingual group with analytic language knowledge; English Language Teaching, English Language and Literature, bilingual group with analytic language knowledge) in terms of their fields of interest in language. The number of the participants was sufficient to be able to generalize the results, with 64, 67 and 77 subjects in each subgroup, respectively, making 208 in totals. Regarding the participants’ ages, they were all roughly the same age and mostly in the third year of university. A total of 175 participants were aged 20 to 22, the typical age of Turkish third-year students. Twenty-four participants ranged in age from 23 to 25, Vestnik_za_tuje_jezike_2022_FINAL.indd 151 Vestnik_za_tuje_jezike_2022_FINAL.indd 151 24. 01. 2023 09:18:49 24. 01. 2023 09:18:49 152 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES six individuals were over 26, and three participants were between the ages of 17 and 19, typical of first-year students. 3.4 Data Analysis The data obtained from the questionnaire and test was examined using IBM’s Statistical Package for the Social Sciences (SPSS) 22. Item analysis of the questions and the reli- ability of the test were examine by using essential the related formulas in Microsoft Ex- cel. The descriptive statistics of all the items, Kruskal-Wallis Test and one-way ANOV A were carried out using IBM SPSS 22. 3.4.1 The Validity of the Test 3.4.1.1 Item Analysis Two hundred and ten participants were ranked from the highest to lowest test scores on the Indonesian test and then divided into two groups. One hundred and five subjects were placed into each group, namely the upper and lower groups, with the former including the participants with higher grades and the latter those with lower grades. The total number of correct answers to each question is shown below. Table 1: The number of correct answers by both groups for each item I1 I2 I3 I4 I5 I6 I7 I8 I9 I10 I11 I12 I13 I14 I15 Total 166 92 153 156 46 118 48 95 63 93 45 39 90 23 104 Upper group 98 58 86 101 41 88 45 73 59 76 38 39 76 23 86 Lower group 68 34 67 55 5 30 3 22 4 17 7 0 14 0 18 As shown in the table above, the upper group answered the fourth item (i.e., What was the advertisement for that Edo sees? Pay attention to the word “pertandingan” (line 6).) correctly 101 times and answered the 14 th item (i.e., The last word of the text is sekolah, what is the meaning of this word?) wrongly 23 times. These are the extreme values for the upper group. As for lower group, they answered the first item (i.e., Who is the main character?) correctly 68 times and could not answer the 12 th (i.e., Guru Bahasa Indonesia masuk ke kelas. Translate this sentence into Turkish (lines 14 and 15).) and 14 th items (i.e., The last word of the text is sekolah, what is the meaning of this word?) correctly at all. Vestnik_za_tuje_jezike_2022_FINAL.indd 152 Vestnik_za_tuje_jezike_2022_FINAL.indd 152 24. 01. 2023 09:18:49 24. 01. 2023 09:18:49 153 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... 3.4.1.2 Item Difficulty The difficulty of each item in the test was calculated, and the results are presented below. Table 2: The difficulty of the items I1 I2 I3 I4 I5 I6 I7 I8 I9 I10 I11 I12 I13 I14 I15 0.79 0.44 0.73 0.74 0.22 0.56 0.23 0.45 0.30 0.44 0.21 0.18 0.43 0.11 0.49 The numbers above indicate that item one i.e. “Who is the main character?” is the easiest item in the test, with a value of .79, and item 14, i.e. “The last word of the text is sekolah, what is the meaning of this word?” is the most difficult question with a value of .11. The remaining 13 items’ values can also be seen in the table. 3.4.1.3 Item Discrimination The discrimination of each question was calculated by using point-biserial correlation, with the results shown below. Table 3: The discrimination of the items I1 I2 I3 I4 I5 I6 I7 I8 I9 I10 I11 I12 I13 I14 I15 0.43 0.33 0.34 0.56 0.56 0.62 0.59 0.53 0.62 0.62 0.52 0.66 0.66 0.56 0.70 From the results above, it can be concluded that the point-biserial correlation coefficients of the items were between 0.33 and 0.70, which means they have high discrimination. Items with values between 0.30 and 0.40 can be used, although some changes may also be made. Items with values higher than 0.40 are highly distinctive and can be directly used, and this is the case with 13 items in the test. The second item in the test i.e. “A lot of names are mentioned in the text, explain what kind of relationship these people have with each other”, is the least distinctive, with a value of 0.33, while the 15 th item i.e., “Indonesian texts often have a strong moralizing message, the writer wants to make some point. What do you think is the moral to this story?”, is the most distinctive in the test. The overall values indicate that all the items in this text are highly distinctive. 3.4.2 The Reliability of the Test Cronbach’s alpha was used in order to evaluate the reliability of the test, and with a value of 0.84 the result shows that the test has a high level of reliability. Vestnik_za_tuje_jezike_2022_FINAL.indd 153 Vestnik_za_tuje_jezike_2022_FINAL.indd 153 24. 01. 2023 09:18:49 24. 01. 2023 09:18:49 154 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES 3.4.3 Checking Assumptions Certain criteria need to be met before the gathered data is analysed, namely the control of the outliers and the test of normality, as explained below. 3.4.3.1 The Control of the Outliers Figure 1: The control of the outliers From the data in Figure 1, it is evident that two participants from the departments of Turkish Language and Literature and English Language Teaching and Literature, respectively, were identified as outliers (113 and 161), and for that reason were both excluded from the study. The one-way analysis of variance (ANOV A) was used to determine whether there were any significant differences between the means of two or more independent groups, and for this reason one-way ANOV A was used as an analysis technique (Field, 2009). Two assumptions must be met in order to use one-way ANOV A, which were the normal distribution of the data and the homogeneity of the variances. 3.4.3.2 The Test of Normality The homogeneity of variances needed to be tested for the normal distribution of the data. Skewness and Kurtosis values were all calculated for each department in order to make sure the homogeneity of variances was warranted or not, and values that were between -1.96 and +1.96 were accepted as part of the normal distribution, as seen in the table below. Vestnik_za_tuje_jezike_2022_FINAL.indd 154 Vestnik_za_tuje_jezike_2022_FINAL.indd 154 24. 01. 2023 09:18:49 24. 01. 2023 09:18:49 155 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... Table 4: Skewness and kurtosis values History & Philosophy Turkish Language and Literature English Language Teaching & English Language and Literature Skewness -0.123 -1.266 0.682 Kurtosis -0.862 -0.432 -1.086 The table above shows that all of the values are acceptable, indicating that a normal distribution is derived. The number of the subjects in each department was over than 50, and thus the Kol- mogorov-Smirnov test of normality was used. With this the values must be lower than 0.001, meaning that the expected values and observed values are equal. The values from the Kolmogorov-Smirnov test of normality can be seen in the table below. Table 5: Test of normality Department Kolmogorov-Smirnov a Statistic df Sig. Test Score History & Philosophy .154 64 .001 Turkish Language and Literature .140 67 .001 English Language Teaching & English Language and Literature .142 77 .001 As shown in the table above, the values are all acceptable ( p= .001) for the use of one- way ANOV A, and the results related to these statistics are examined in the following section. 4 RESULTS To determine whether there are any statistically significant differences between the means of the three study groups, one-way ANOV A and the Kruskal-Wallis test were performed. The results of one-way ANOV A are described and analysed below only if the value of homogeneity was higher than 0.05, which means the homogeneity is at an acceptable level and the one-way ANOV A test can be applied. Otherwise, the Kruskal-Wallis Test was carried out as a non-parametric test. Vestnik_za_tuje_jezike_2022_FINAL.indd 155 Vestnik_za_tuje_jezike_2022_FINAL.indd 155 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 156 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES 4.1 The Effect of Being Sequential Bilingual on Metalinguistic Awareness Table 6: Descriptive statistics about students’ departments and test scores Test Score N Mean Std. Deviation Std. Error Minimum Maximum History & Philosophy 64 4.28 2.257 .282 0 9 Turkish Language and Literature 67 3.72 1.765 .216 0 7 English Language Teaching & English Language and Literature 77 10.32 2.250 .256 5 15 Total 208 6.34 3.721 .258 0 15 It can be seen in the table above that the participants from History & Philosophy and Turkish Language and Literature had the lowest scores on the test of 0 and highest scores of 9 and 7, respectively, which are close to each other. However, the participants from English Language Teaching & English Language and Literature had the lowest score of 5 and highest score of 15 (i.e., some participants answered all the questions correctly). Students from English Language departments were regarded as sequential bilinguals, and their superiority over the other two groups is obvious in terms of test scores. For example, no participant from the other two groups could answer all the questions correctly, while no student with an English language background answered all the questions incorrectly. Table 7: Test of homogeneity of variances (among departments) Test Score Levene Statistic df1 df2 Sig. 2.197 2 205 .114 The test of homogeneity of variances was not significant. The results revealed that the assumption of equal variances was met (p= .114), so the results of the one-way ANO- V A can be analysed and discussed, with the results shown in the following table. Vestnik_za_tuje_jezike_2022_FINAL.indd 156 Vestnik_za_tuje_jezike_2022_FINAL.indd 156 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 157 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... Table 8: One-way ANOV A (among departments) Test Score Sum of Squares df Mean Square F Sig. Between Groups 1955.010 2 977.505 219.861 .000 Within Groups 911.433 205 4.446 Total 2866.442 207 As shown in the table above, the f-value is calculated as 219.861, the significance level is .000 (< 0.05), and there is a significant difference between the groups. However, the results show that the effects of the groups cannot be seen among each other, and in order to evaluate the relationships within the groups a multiple comparisons test (i.e., Tukey HSD) is carried out, with the results shown below. Table 9: Multiple comparisons (among departments) Dependent Variable: Test Score Tukey HSD (I) Department (J) Department Mean Difference (I-J) Std. Error Sig. History & Philosophy Turkish Language and Literature .565 .369 .278 English Language Teaching & Literature -6.043 * .357 .000 Turkish Language and Literature History & Philosophy -.565 .369 .278 English Language Teaching & Literature -6.608 * .352 .000 English Language Teaching & Literature History & Philosophy 6.043 * .357 .000 Turkish Language and Literature 6.608 * .352 .000 * The mean difference is significant at the 0.05 level. As the table shows, the participants from English Language Teaching & English Language and Literature performed significantly better when compared to those from History & Philosophy and Turkish Language and Literature (p = .000) on the ILT. This finding also answered the question raised by this study with regard to sequential bilinguals and monolinguals, with the former group performing better. There was no significant dif- ference between the students from the History & Philosophy and Turkish Language and Vestnik_za_tuje_jezike_2022_FINAL.indd 157 Vestnik_za_tuje_jezike_2022_FINAL.indd 157 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 158 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES Literature departments. The students from the Turkish Language and Literature depart- ment also studied a language, but it was their mother tongue not a foreign language. This may explain why they did not perform better than the History & Philosophy students on the metalinguistic awareness test. 4.2 The Importance of Gender and its Effect on Metalinguistic Awareness As indicated below, the number of the female participants was 162, while there were 46 male participants. Table 10: Gender of the participants Frequency Percent Valid Percenwt Cumulative Percent Valid female 162 77.9 77.9 77.9 male 46 22.1 22.1 100.0 Total 208 100.0 100.0 As for the effects of gender, there was no significant difference between male and female participants, as shown in the tables below. The first table shows the Levene’s Test statistics, and the second table shows the results of one-way ANOV A. Table 11: Test of homogeneity of variances (gender) Test Score Levene Statistic df1 df2 Sig. .141 1 206 .708 The test of homogeneity of variances was not significant. The significance value for Levene’s Test statistics (p= .708) meets the criteria for one-way ANOV A (i.e., the homo- geneity of variances was ensured by the test), as shown in the table below. Table 12: One-way ANOV A (gender) Test Score Sum of Squares df Mean Square F Sig. Between Groups 33,185 1 33,185 2,413 ,122 Within Groups 2833,257 206 13,754 Total 2866,442 207 Vestnik_za_tuje_jezike_2022_FINAL.indd 158 Vestnik_za_tuje_jezike_2022_FINAL.indd 158 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 159 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... The F-value is calculated as 2.413 and the significance value is .112, meaning that there is no significant difference between the male and female participants in the study. It can be inferred from this that gender did not have a role in the level of metalinguistic awareness of the participants. 4.3 The Effect of Age on Metalinguistic Awareness The age range of the participants and their test scores in these groups are shown in the table below. Table 13: Descriptive statistics about students’ ages and test scores Test Score N Mean Std. Deviation Std. Error Minimum Maximum 17-19 3 10.33 5.508 3.180 4 14 20-22 175 6.35 3.752 .284 0 15 23-25 24 6.00 3.176 .648 1 13 26-30 6 5.17 3.601 1.470 1 9 Total 208 6.34 3.721 .258 0 15 Most of the students were in 20-22 age range, and this group had a highest score of 15, but also a lowest score of 0. The students in the 17-19 age range had the highest mean value of 10.33, with three participants, while the 26-30 age range had the lowest mean value of 5.17, with six subjects. Table 14: Test of homogeneity of variances for the age of the participants Test Score Levene Statistic df1 df2 Sig. 1.327 3 204 .267 The test of the homogeneity of variances was not significant, with the results show- ing equal variances (p= .267), and the one-way ANOV A results shown in the following table. Vestnik_za_tuje_jezike_2022_FINAL.indd 159 Vestnik_za_tuje_jezike_2022_FINAL.indd 159 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 160 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES Table 15: One-way ANOV A (age) Test Score Sum of Squares df Mean Square F Sig. Between Groups 58.908 3 19.636 1.427 .236 Within Groups 2807.534 204 13.762 Total 2866.442 207 From the results shown in the table above, it can be inferred that the age of the par- ticipants had no influence on the results of the metalinguistic awareness test (p= .236). 4.4 The Effect of the Age Students Started Language Learning on Metalinguistic Awareness Statistical data with regard to the age the students started language learning and their test scores are shown in the table below. Table 16: Descriptive statistics about students’ language learning starting age Test Score N Mean Std. Deviation Std. Error Minimum Maximum 0-6 14 4.43 3.975 1.062 0 14 7-13 154 6.79 3.715 .299 0 15 14-18 33 5.58 3.437 .598 1 13 19-24 4 4.00 2.449 1.225 1 6 Total 205 6.38 3.729 .260 0 15 As shown in the table above, most of the subjects started language learning between the ages of 7 and 13, and they had the highest mean score of 6.79. Only four participants stated that they started language learning between 19 and 24, they also had the lowest mean score of 4.00. Table 17: Test of homogeneity of variances about students’ language learning starting age Test Score Levene Statistic df1 df2 Sig. .822 3 201 .483 Vestnik_za_tuje_jezike_2022_FINAL.indd 160 Vestnik_za_tuje_jezike_2022_FINAL.indd 160 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 161 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... The test of homogeneity test of variances was not significant, and the results showed that the assumption of equal variances was met (p= .483), with the one-way ANOV A results presented in the following table. Table 18: One-way ANOV A for students’ language learning starting age Test Score Sum of Squares df Mean Square F Sig. Between Groups 123.482 3 41.161 3.050 .030 Within Groups 2712.840 201 13.497 Total 2836.322 204 Although the table above shows an important relationship in terms of students’ lan- guage learning starting age, the multiple comparisons shown in the table below do not indicate a significant relationship within the groups. The F test examines whether all four means are equal, and the alternative that is frequently used – that at least one pair of means differs – is not quite accurate. The better alternative is that there is a linear contrast among the four means that is significantly different from zero. A pairwise difference is one example of a linear contrast, but there are other linear contrasts that the Tukey test does not look at. For example, it might be that the first mean does not differ significantly from the third mean, and the second mean does not differ significantly from the fourth mean, but maybe an average of the first and second means differs significantly from an average of the third and fourth means. Or maybe the fourth mean is slightly smaller than the other means, but not enough to be statistically significant for any pair. Table 19: Multiple comparisons about students’ language learning starting age Dependent Variable: Test Score Tukey HSD (I) At what age did you start learning English? (J) At what age did you start learning English? Mean Difference (I-J) Std. Error Sig. 0-6 7-13 -2.364 1.026 .100 14-18 -1.147 1.172 .762 19-24 .429 2.083 .997 7-13 0-6 2.364 1.026 .100 14-18 1.216 .705 .313 19-24 2.792 1.861 .439 Vestnik_za_tuje_jezike_2022_FINAL.indd 161 Vestnik_za_tuje_jezike_2022_FINAL.indd 161 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 162 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES (I) At what age did you start learning English? (J) At what age did you start learning English? Mean Difference (I-J) Std. Error Sig. 14-18 0-6 1.147 1.172 .762 7-13 -1.216 .705 .313 19-24 1.576 1.945 .850 19-24 0-6 -.429 2.083 .997 7-13 -2.792 1.861 .439 14-18 -1.576 1.945 .850 The previous table (i.e., table 18) showed some significant results, but no significant relationship can be observed in table 19 for the reasons mentioned above, i.e. that the significance can be about the mean of different groups, not just that of one specific group. 4.5 The Effect of the Length of Language Learning on Metalinguistic Awareness The table below shows the statistics about students’ language learning backgrounds in terms of years, and this will be compared to their test scores. Table 20: Descriptive statistics about the length of language learning Test Score N Mean Std. Deviation Std. Error Minimum Maximum 0-2 4 5.25 2.217 1.109 3 8 3-5 16 3.44 2.732 .683 1 12 6-8 44 5.27 2.748 .414 0 12 9-11 72 6.03 3.685 .434 0 14 12-14 62 8.40 3.835 .487 0 15 15-17 2 8.50 .707 .500 8 9 Total 200 6.40 3.739 .264 0 15 The table indicates that the students with 15-17 years of language learning back- ground had the highest mean of 8.50, and the 3-5 years of language learning background group had the lowest mean score, 3.44. Vestnik_za_tuje_jezike_2022_FINAL.indd 162 Vestnik_za_tuje_jezike_2022_FINAL.indd 162 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 163 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... Table 21: Test of homogeneity of variances about the length of language learning Test Score Levene Statistic df1 df2 Sig. 2.848 5 194 .017 The result of the homogeneity test of variances was significant (p= .017), and did not confirm the assumption of equal variances, and thus the one-way ANOV A test was not used. In contrast, the independent samples Kruskal-Wallis test was applied, since it is a non-parametric test, and the results are shown below. Table 22: Summary of hypothesis testing about the length of language learning Null Hypothesis Test Sig. Decision 1 The distribution of test scores is the same across categories of responses to “For how long have you been learning English?” Independent- Samples Kruskal- Wallis Test .000 Reject the null hypothesis. Asymptotic significances are shown. The significance level is .05. As shown in the table above, the null hypothesis is rejected, meaning that there is a significant relationship among the items (p= .000). The table below indicates the relation- ship among items and compares them with each other. Table 23: Multiple comparisons about the length of language learning Dependent Variable: Test Score Tukey HSD (I) For how long have you been learning English? (J) For how long have you been learning English? Mean Difference (I-J) Std. Error Sig. 0-2 3-5 1.813 1.930 .936 6-8 -.023 1.803 1.000 9-11 -.778 1.774 .998 12-14 -3.153 1.781 .487 15-17 -3.250 2.990 .886 Vestnik_za_tuje_jezike_2022_FINAL.indd 163 Vestnik_za_tuje_jezike_2022_FINAL.indd 163 24. 01. 2023 09:18:50 24. 01. 2023 09:18:50 164 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES (I) For how long have you been learning English? (J) For how long have you been learning English? Mean Difference (I-J) Std. Error Sig. 3-5 0-2 -1.813 1.930 .936 6-8 -1.835 1.008 .455 9-11 -2.590 .954 .077 12-14 -4.966 * .968 .000 15-17 -5.063 2.590 .372 6-8 0-2 .023 1.803 1.000 3-5 1.835 1.008 .455 9-11 -.755 .661 .863 12-14 -3.130 * .681 .000 15-17 -3.227 2.496 .789 9-11 0-2 .778 1.774 .998 3-5 2.590 .954 .077 6-8 .755 .661 .863 12-14 -2.375 * .598 .001 15-17 -2.472 2.475 .918 12-14 0-2 3.153 1.781 .487 3-5 4.966 * .968 .000 6-8 3.130 * .681 .000 9-11 2.375 * .598 .001 15-17 -.097 2.481 1.000 15-17 0-2 3.250 2.990 .886 3-5 5.063 2.590 .372 6-8 3.227 2.496 .789 9-11 2.472 2.475 .918 12-14 .097 2.481 1.000 * The mean difference is significant at the 0.05 level. The table above shows that participants who had been learning English language for 12-14 years performed significantly better than those who had been doing so for 3-5 years, 6-8 years, or 9-11 years. Having a longer language learning background thus had positive effects the results of the on metalinguistic awareness test. No significant relation- ship was found among the other groups. Vestnik_za_tuje_jezike_2022_FINAL.indd 164 Vestnik_za_tuje_jezike_2022_FINAL.indd 164 24. 01. 2023 09:18:51 24. 01. 2023 09:18:51 165 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... 4.6 The Effect of a Preparatory Class of English on Metalinguistic Awareness The following table shows the number of the participants who had taken a preparatory class in English and their means, standard deviations, minimum and maximum scores. Table 24: Descriptive statistics for a preparatory class of English Test Score N Mean Std. Deviation Std. Error Minimum Maximum High School 9 4.78 2.774 .925 0 9 University 45 10.29 2.677 .399 2 15 None 154 5.27 3.224 .260 0 14 Total 208 6.34 3.721 .258 0 15 There were 45 subjects who had taken a preparatory English class at university level, while nine had done so in high school. The other 154 participants, the majority in this study, stated that they had no preparatory English classes in their backgrounds. Table 25: Test of homogeneity of variances with regard to preparatory English classes Test Score Levene Statistic df1 df2 Sig. 1,135 2 205 .323 The result of the test of homogeneity test of variances was not significant, and con- firmed that the criteria of equal variances was met (p= .323). The results of the subse- quent one-way ANOV A test are shown below. Table 26: One-way ANOV A with regard to preparatory English classes Test Score Sum of Squares Df Mean Square F Sig. Between Groups 899.097 2 449.548 46.844 .000 Within Groups 1967.345 205 9.597 Total 2866.442 207 The results of one-way ANOV A are statistically significant (p= .000 level), and the Tukey HSD was thus applied as a post hoc test in order to find out between which groups there is a significant difference, with the results analysed below. Vestnik_za_tuje_jezike_2022_FINAL.indd 165 Vestnik_za_tuje_jezike_2022_FINAL.indd 165 24. 01. 2023 09:18:51 24. 01. 2023 09:18:51 166 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES Table 27: Multiple comparisons with regard to a preparatory English class Dependent Variable: Test Score Tukey HSD (I) Have you studied at a preparatory school of English? (J) Have you studied at a preparatory school of English? Mean Difference (I-J) Std. Error Sig. High School University -5.511 * 1.131 .000 None -.495 1.062 .887 University High School 5.511 * 1.131 .000 None 5.016 * .525 .000 None High School .495 1.062 .887 University -5.016 * .525 .000 *. The mean difference is significant at the 0.05 level. As can be seen in the table, the participants who had a preparatory English class at university performed significantly better than both the participants who had such a class at high school and those who had never taken a preparatory class (p= .000). It can thus be deduced that taking a preparatory English class at university enhances learners’ metalin- guistic awareness skills. 4.7 The Effect of the Proficiency Level of the Students on Metalinguistic Awareness The self-reported proficiency level of the participants was also asked, and their answers can be seen in the table below. Table 28: Descriptive statistics for the proficiency level of the participants Test Score N Mean Std. Deviation Std. Error Minimum Maximum Very poor 25 3.28 2.031 .406 0 7 Poor 54 4.17 2.196 .299 0 9 Acceptable 52 5.38 2.911 .404 0 13 Good 62 9.74 3.002 .381 2 15 Very good 11 9.00 4.195 1.265 0 15 Total 204 6.32 3.743 .262 0 15 Vestnik_za_tuje_jezike_2022_FINAL.indd 166 Vestnik_za_tuje_jezike_2022_FINAL.indd 166 24. 01. 2023 09:18:51 24. 01. 2023 09:18:51 167 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... Four participants did not state their proficiency levels, while the answers of the re- maining 204 participants, and their mean, standard deviation, minimum and maximum values, are shown in the table above. Table 29: Test of homogeneity of variances about the proficiency levels of the participants Test Score Levene Statistic df1 df2 Sig. 2,319 4 199 .058 As shown in the table above, the result of the test of homogeneity of variances was not significant (p= .058), and thus the assumption of equal variances was met and one- way ANOV A could be applied, with the results shown in the table below. Table 30: One-way ANOV A about the proficiency levels of the participants Test Score Sum of Squares df Mean Square F Sig. Between Groups 1331,928 4 332,982 43,804 ,000 Within Groups 1512,719 199 7,602 Total 2844,647 203 The table above shows that there is a significant difference between groups in terms of language proficiency (p= .000). Tukey HSD is thus carried out so as to understand the relationships among the groups, with the results shown below. Table 31: Multiple comparisons about the proficiency levels of the participants Dependent Variable: Test Score Tukey HSD (I) What is your English language proficiency? (J) What is your English language proficiency? Mean Difference (I-J) Std. Error Sig. Very poor Poor -.887 .667 .673 Acceptable -2.105 * .671 .017 Good -6.462 * .653 .000 Very good -5.720 * .998 .000 Vestnik_za_tuje_jezike_2022_FINAL.indd 167 Vestnik_za_tuje_jezike_2022_FINAL.indd 167 24. 01. 2023 09:18:51 24. 01. 2023 09:18:51 168 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES (I) What is your English language proficiency? (J) What is your English language proficiency? Mean Difference (I-J) Std. Error Sig. Poor Very poor .887 .667 .673 Acceptable -1.218 .536 .158 Good -5.575 * .513 .000 Very good -4.833 * .912 .000 Acceptable Very poor 2.105 * .671 .017 Poor 1.218 .536 .158 Good -4.357 * .518 .000 Very good -3.615 * .915 .001 Good Very poor 6.462 * .653 .000 Poor 5.575 * .513 .000 Acceptable 4.357 * .518 .000 Very good .742 .902 .923 Very good Very poor 5.720 * .998 .000 Poor 4.833 * .912 .000 Acceptable 3.615 * .915 .001 Good -.742 .902 .923 * The mean difference is significant at the 0.05 level. According to Table 31, the participants who stated their proficiency as “good” and “very good” performed significantly better than those who declared their proficiency as “very poor”, “poor”, and “acceptable” (p= .000 and p= .001). There is no significant dif- ference with regard to the effect of the proficiency level of the participants who stated their levels “good” and “very good” in terms of metalinguistic awareness. 4.8 The Effect of Languages Being Spoken Other than Turkish or English on Metalinguistic Awareness Languages other than Turkish or English are spoken by some of the participants, as shown in the following table. Table 32: Descriptive statistics for speaking a language other than Turkish and English Test Score N Mean Std. Deviation Std. Error Minimum Maximum No 190 6.38 3.632 .263 0 15 Yes 18 5.89 4.664 1.099 0 14 Total 208 6.34 3.721 .258 0 15 Vestnik_za_tuje_jezike_2022_FINAL.indd 168 Vestnik_za_tuje_jezike_2022_FINAL.indd 168 24. 01. 2023 09:18:52 24. 01. 2023 09:18:52 169 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... As shown in the table above, only 18 participants said “Yes” to this item, but a test of homogeneity and one-way ANOV A were also carried out, as shown in the tables below. Table 33: Test of homogeneity of variances about languages spoken other Turkish and English Test Score Levene Statistic df1 df2 Sig. 1,344 1 206 .248 The table above indicates that the test of homogeneity of variances was not signifi- cant, and confirmed the assumption of equal variances (p= .248). As for the research questions examined in this study, the first question was about the relationship between metalinguistic awareness and analytic language knowledge. The re- sults show that analytic language knowledge – especially if it belongs to a second lan - guage (i.e., not the mother tongue) – has a significant effect on metalinguistic awareness. The second question focused on the departments of the students, and the results showed that the students from English Language History, Philosophy and Turkish Language and Literature. These last two groups Teaching and English Language and Literature depart- ments performed significantly better than those from the departments of had no significant differences in terms of metalinguistic awareness. The third research question dealt with the information gathered from the demographic questionnaire. As the primary objective of the study was to find out the possible relationship between analytic language knowledge and metalinguistic awareness, the effects of gender, age, starting age of language learn - ing, previous years of language learning, English preparatory classes, proficiency level and speaking languages other than Turkish or English were analysed. Among these variables, attending English preparatory classes, proficiency level, and the length of language learning all had significant effects on the results of the metalinguistic awareness test. Students with preparatory English classes (at university), a higher proficiency level (i.e., as “very good” and “good” depending on the participants’ perception and when compared to the lower pro - ficiency levels of “very poor”, “poor”, and “acceptable”), and having at least 12-14 years of language learning background (i.e., when compared to 3-5, 6-8, and 9-11 years) performed significantly better than the other group or groups in each section. 5 DISCUSSION AND CONCLUSION The primary objective of this study was to clarify the relationship between analytic lan - guage knowledge and metalinguistic awareness. For this reason, three groups of students were chosen from the Turkish Language and Literature, History and Philosophy, English Language Teaching and English Language and Literature departments. The subjects from Vestnik_za_tuje_jezike_2022_FINAL.indd 169 Vestnik_za_tuje_jezike_2022_FINAL.indd 169 24. 01. 2023 09:18:52 24. 01. 2023 09:18:52 170 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES English language departments were regarded as sequential bilinguals, as they did not learn English at an early age. The participants from the Turkish Language Department were mostly monolinguals studying the language in detail. The students of the History and Phi- losophy departments did not study any language in detail, and they were selected to clarify whether any group had superiority over the others in the metalinguistic awareness task. The results showed that being able to speak a language other than one’s mother tongue had a significant effect on the process of learning new languages and on meta- linguistic awareness tests. To put it another way, the results of this study proved that the level of metalinguistic awareness and the number of the languages that one knows are significantly related to each other. The students from English Language Teaching and English Language and Literature departments outperformed the other two groups at a significant level (p= .000) on the metalinguistic awareness test, and thus sequential bi- linguals were better than the other groups. This result supports the idea that being able to speak at least two languages at a high proficiency level has a significant impact on learn- ing new languages (Ricciardelli, 1992). Turning to the results of similar studies, in parallel with the findings of this study Cummins (1978) reported that bilingual students performed significantly better than monolinguals in tasks related to the awareness of arbitrary nature of the word referent relationship. Moreover, Lauchlan, Parisi and Fadda (2012) also found similar results, as bilinguals scored better on cognitive control, problem-solving tasks, metalinguistic awareness and working memory tests, as in the current study. However, there was no significant relationship between Turkish Language and Literature students and History and Philosophy students, although the latter scored better than the former. Therefore, as it clearly seen in this study, balanced bilinguals had a significant advantage over their monolingual counterparts, and bilinguals can benefit whether they are consciously aware of these advantages or not. Moreover, it was also found out that attending a preparatory English class and a stu - dent’s proficiency level were the other two variables that yielded significant relationships, because those students that had a preparatory class scored significantly better than those that did not (p= .000). As for proficiency levels, the students stating that they had “good” or “very good” English performed significantly better than all the other groups ( p= .000 and p= .001). However, it should also be kept in mind that the reason for the significant results in those two groups might be due to the fact that most of the students in those groups were from the English Language Teaching and English Language and Literature departments. Moreover, students who had preparatory English classes also stated that they had a high level of English proficiency. On the other hand, there was no significant relationship between the students from the History, Philosophy and Turkish Language and Literature departments. Demographic data was also gathered, including age and gender, but the overall re- sults showed that the department of the participants and the test scores were the most important variables, as no significant differences were found among the other variables. Vestnik_za_tuje_jezike_2022_FINAL.indd 170 Vestnik_za_tuje_jezike_2022_FINAL.indd 170 24. 01. 2023 09:18:52 24. 01. 2023 09:18:52 171 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... Finally, as the present findings empirically suggest that sequential bilinguals have higher levels of metalinguistic awareness at the tertiary level, different curricula for dif- ferent departments should be used in terms of additional language courses. For instance, the students from English language backgrounds can have additional language courses that take their metalinguistic abilities into account. Moreover, the lessons can be pre- sented in a more explicit way, since such analytic knowledge helps to improve language learning skills and so accelerate the process (Woll, 2019). Last but not the least, it is hoped that the findings of this study will contribute to the growing literature on the relationship between analytic language knowledge and metalin- guistic awareness while learning a new language, and that foreign language teachers and researchers in the field will benefit from the results in terms of the effects of that analytic language knowledge has on metalinguistic awareness. BIBLIOGRAPHY ATILGAN, Hakan/Adnan KAN/Nuri DOĞAN (2007) Eğitimde Ölçme ve Değerlendirme. (H. Atılgan, Ed.). Ankara, Turkey: Anı Yayıncılık. BARAC, Raluca/Ellen BIALYSTOK (2011) Cognitive development of bilingual children. Language Teaching, 44(01), 36–54. http://doi.org/10.1017/S0261444810000339 BERKEN, Jonathan A./Vincent L. GRACCO/Denise KLEIN (2017) Early bilingualism, language attainment, and brain development. Neuropsychologia, 98, 220–227. https://doi.org/10.1016/j.neuropsychologia.2016.08.031 BIALYSTOK, Ellen/Ellen B. RY AN (1985) Toward a definition of metalinguistic skill. Merrill-Palmer Quarterly, 31(3), 229–251. BIALYSTOK, Ellen (2003) Bilingualism in Development: Language, Literacy,and Cog- nition. Cambridge: Cambridge University Press. BROWN, James D./Theodore S. RODGERS (2003) Doing Second Language Research. Oxford, UK: Oxford University Press. CENOZ, Jasone (2003) The additive effect of bilingualism on third language acquisition: A review. International Journal of Bilingualism, 7(1), 71–87. http://doi.org/10.117 7/13670069030070010501 CUMMINS, James (1976) The influence of bilingualism on cognitive growth: A synthe- sis of research findings and explanatory hypotheses. Working Papers on Bilingual- ism, (9), 121–129. CUMMINS, James (1978) Bilingualism and the Development of Metalinguistic Awareness. Journal of Cross-Cultural Psychology, 9(2), 131–149. http://doi. org/10.1177/002202217892001 CUMMINS, James (1979) Linguistic interdependence and the educational development of bilingual children. Review of Educational Research, 49(2), 222–251. http://doi. org/10.3102/00346543049002222 Vestnik_za_tuje_jezike_2022_FINAL.indd 171 Vestnik_za_tuje_jezike_2022_FINAL.indd 171 24. 01. 2023 09:18:52 24. 01. 2023 09:18:52 172 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES FALK, Ylva/Christina LINDQVIST/Camilla BARDEL (2013) The role of L1 explicit metalinguistic knowledge in L3 oral production at the initial state. Bilingualism: Language and Cognition, 18(2), 1–9. http://doi.org/10.1017/S1366728913000552 GONZALVES, Lisa (2020) Development and demonstration of metalinguistic awarene- ss in adult ESL learners with emergent literacy. Language Awareness, 30(2), 134- 151. https://doi.org/10.1080/09658416.2020.1776721 HAKUTA, Kenji/Judith SUBEN (1985) Bilingualism and cognitive development. An- nual Review of Applied Linguistics, 6, 35–45. HAKUTA, Kenji (1986) Bilingualism and intelligence. In MIRROR OF LANGUAGE The Debate on Bilingualism (pp. 14–44). New York: Basic Books. HAKUTA, Kenji (1987) Degree of bilingualism and cognitive ability in mainland Puerto Rican children. Child Development, 58(5), 1372–1388. http://doi.org/10.1111/1467- 8624.ep8591435 HAKUTA, Kenji/Eugene E. GARCIA (1989) Bilingualism and education. American Psychologist, 44(2), 374–379. http://doi.org/10.1037//0003-066X.44.2.374 HU, Guangwei (2011) Metalinguistic knowledge, metalanguage, and their relationship in L2 learners. System, 39(1), 63–77. http://doi.org/10.1016/j.system.2011.01.011 KOPEČKOVÁ, Romana (2018) Exploring metalinguistic awareness in L3 phonological acquisition: the case of young instructed learners of Spanish in Germany. Language Awareness, 27(1-2), 153-166. https://doi.org/10.1080/09658416.2018.1432629. LAUCHLAN, Fraser/Marinella PARISI/Roberta FADDA (2012) Bilingualism in Sardinia and Scotland: Exploring the cognitive benefits of speaking a “minor- ity” language. International Journal of Bilingualism, 17(1), 43–56. http://doi. org/10.1177/1367006911429622 MACKEY, Alison/Susan M. GASS (2005) Second Language Research: Methodology and Design. New Jersey: Lawrence Erlbaum Associates. MERTZ, Elizabeth/Jonathan YOVEL (2009) Metalinguistic Awareness. In Cognition and Pragmatics (pp. 250–271). Amsterdam/Philadelphia: John Benjamins Pub- lishing Company. PARADIS, Michel (2004) A Neurolinguistic Theory of Bilingualism. Amsterdam/Phila- delphia: John Benjamins Publishing Company. http://doi.org/10.1075/sibil.18 RICCIARDELLI, Lina A. (1992) Bilingualism and cognitive development in relation to threshold theory. Journal of Psycholinguistic Research, 21(4), 301–316. http://doi. org/10.1007/BF01067515 ROEHR, Karen (2007) Metalinguistic knowledge and language ability in university-lev- el L2 learners. Applied Linguistics, 29(2), 173–199. http://doi.org/10.1093/applin/ amm037 ROMAINE, Suzanne (2013) The bilingual and multilingual community. In Bhatia, T. K./W. C. Ritchie (Eds.), The Handbook of Bilingualism and Multilingualism (2nd ed., pp. 445–465). Wiley-Blackwell. Vestnik_za_tuje_jezike_2022_FINAL.indd 172 Vestnik_za_tuje_jezike_2022_FINAL.indd 172 24. 01. 2023 09:18:52 24. 01. 2023 09:18:52 173 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... TER KUILE, Hagar/Michiel VELDHUIS/Suzanne C. V AN VEEN/Jelte M. WICHERTS (2011) Bilingual education, metalinguistic awareness, and the understanding of an unknown language. Bilingualism: Language and Cognition, 14(02), 233–242. http://doi.org/10.1017/S1366728910000258 WOLL, Nina (2019) How French speakers reflect on their language: A critical look at the concept of metalinguistic awareness. Language Awareness, 28(1), 49-73. https://doi. org/10.1080/09658416.2019.1567743 POVZETEK PRIMERJA V A RAZMERJA MED ANALITIČNIM JEZIKOVNIM ZNANJEM IN META- JEZIKOVNIM ZA VEDANJEM PRI ENOJEZIČNIH IN ZAPOREDNO DVOJEZIČNIH ŠTUDENTIH Analitično jezikovno znanje je pogosto predmet strokovnih razprav, njegov morebitni vpliv na metajezikovno zavedanje pa vzbuja veliko zanimanja med raziskovalci. Namen naše raziskave je bil proučiti povezavo med analitičnim jezikovnim znanjem in metajezikovnim zavedanjem. V raziskavo smo vključili 210 študentov treh javnih univerz v Turčiji, ki obiskujejo eno od naslednjih smeri študija: zgodovina in filozofija (enojezični študenti brez analitičnega jezikovnega znanja), turški jezik in književnost (enojezični študenti, ki študirajo oz. analitično proučujejo turščino, svoj materni jezik) ter poučevanje angleščine, angleški jezik in književnost (zaporedno dvojezični štu- denti, ki študirajo oz. analitično proučujejo angleščino, svoj drugi jezik). Sodelujoči v raziskavi so morali izpolniti vprašalnik o preteklih jezikovnih izkušnjah in opraviti test (prirejen po Ter Kuile et al., 2011), napisan v indonezijskem jeziku, da bi ocenili raven svojega metajezikovnega zavedanja. Z vprašalniki pridobljene podatke smo analizirali s pomočjo programa IBM SPSS 22, nato pa smo primerjali rezultate študentov različnih študijskih smeri. Analiza je pokazala, da so se študenti pedagoške smeri angleščine in smeri angleški jezika in književnost odrezali bistveno bolje od študentov drugih dveh študijskih smeri. Med študenti zgodovine in filozofije ter študenti turškega jezika in književnosti nismo zaznali večjih razlik. Rezultati empirične raziskave kažejo, da imajo zaporedno dvojezični posamezniki višjo stopnjo metajezikovnega zavedanja, zato bi bilo učne načrte različnih študijskih smeri smiselno prenoviti in v njih predvideti dodatne jezikovne tečaje. Večjezičnim študentom bi ob upoštevanju njihovih metajezikovnih zmožnosti lahko omo- gočili obiskovanje dodatnih jezikovnih izobraževanj. Ključne besede: metajezikovno zavedanje, zaporedno dvojezični posamezniki, analitično jezi- kovno znanje Vestnik_za_tuje_jezike_2022_FINAL.indd 173 Vestnik_za_tuje_jezike_2022_FINAL.indd 173 24. 01. 2023 09:18:52 24. 01. 2023 09:18:52 174 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES ABSTRACT A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL TERTIARY LEVEL STUDENTS ON THE RELATIONSHIP BETWEEN ANALYTIC LANGUAGE KNOWLEDGE AND METALINGUISTIC A W ARENESS* Analytic language knowledge has often been highlighted among experts in the field, and its poten- tial effects on metalinguistic awareness have been of great interest to researchers. This study aims to find the relationship between analytic language knowledge and its potential effect on metalin- guistic awareness. A total of 210 tertiary level students from three different public universities in Turkey participated the study. The subjects were studying in three different departments: History and Philosophy (i.e., with no analytic language abilities and monolinguals), Turkish Language and Literature (i.e., studying Turkish, their mother tongue, analytically and monolinguals), and Eng- lish Language Teaching and English Language and Literature (i.e., studying English, their second language, analytically and sequential bilinguals). The participants were asked to complete a demo- graphic questionnaire and a test (adapted from Ter Kuile et al., 2011) written in Indonesian in order to evaluate their metalinguistic awareness, and the results were compared among the departments. The data gathered through the questionnaires were analysed using IBM SPSS 22. The results show that the participants from English Language Teaching and English Language and Literature per- formed significantly better than the other two groups. No significant difference was found between the History, Philosophy and Turkish Language and Literature departments. The present findings empirically revealed that sequential bilinguals have a greater degree of metalinguistic awareness, and thus alternative curricula for different departments may be adopted in terms of supplementary language courses, and this is one of the implications of this study for future research. Students who have a background in multiple languages, for instance, may be able to enrol in additional language lessons under a plan that is developed by taking metalinguistic abilities into account. Keywords: metalinguistic awareness, sequential bilinguals, analytic language knowledge Vestnik_za_tuje_jezike_2022_FINAL.indd 174 Vestnik_za_tuje_jezike_2022_FINAL.indd 174 24. 01. 2023 09:18:52 24. 01. 2023 09:18:52 175 Birtan Baytar, İsmail Çakır: A COMPARISON OF MONOLINGUAL AND SEQUENTIAL BILINGUAL ... APPENDIX METALINGUISTIC A W ARENESS TEST Vestnik_za_tuje_jezike_2022_FINAL.indd 175 Vestnik_za_tuje_jezike_2022_FINAL.indd 175 24. 01. 2023 09:18:53 24. 01. 2023 09:18:53 176 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES Vestnik_za_tuje_jezike_2022_FINAL.indd 176 Vestnik_za_tuje_jezike_2022_FINAL.indd 176 24. 01. 2023 09:18:53 24. 01. 2023 09:18:53