Available online at www.jlls.org JOURNAL OF LANGUAGE AND LINGUISTIC STUDIES ISSN: 1305-578X Journal of Language and Linguistic Studies, 13(1), 215-231; 2017 An investigation of read speech of Arabic students learning Turkish as a second language in terms of stress and pause Serdar Dermana, Mehmet Bardakçıb*, Mustafa Serkan Öztürkc aGaziantep University, Gaziantep, 27310,Turkey bGaziantep University, Gaziantep, 27310, Turkey bNecmettin Erbakan University, Konya, 42090, Turkey APA Citation: Derman, S., Bardakçı, M. &Öztürk, M.S. (2017). An investigation of read speech of Arabic students learning Turkish as a second language in terms of stress and pause.Journal of Language and Linguistic Studies, 13(1), 215-231. Submission Date:16/01/2017 Acceptance Date:21/02/2017 Abstract Suprasegmental features are essential in conveying meaning; however, they are one of the neglected topics in teaching Turkish as a foreign/second language. This paper aims to examine read speech by Arabic students learning Turkish as a second language and describe their read speech in terms of stress and pause. Within this framework, 34 Syrian students enrolled in Gaziantep University Turkish Teaching Application and Research Center (TÖMER) at B2 level were asked to read a text chosen from one of the books prepared in accordance with the European language portfolio B2 level. Voice recordings were analyzed using Praat and Cool Edit software programs. Students’ stress and pause durations were compared and contrasted according to the criteria set by experts. The results showed that students’ stress and pause patterns showed statistically significant deviations from those criteria. © 2017 JLLS and the Authors - Published by JLLS. Keywords:Turkish as a foreign/second language; suprasegmentals; stress; pause 1. Introduction English is the most popular language for people who learn a foreign language (FL); similarly, Turkish language also gradually gets its significant place in learning as a FL. During the last 30 years, language specialists studied on suprasegmentals instead of segmentals to enhance oral communication (Hismanoglu, 2012; Çelik, 2001; Avery & Ehrlich, 1992; Morley, 1991). For this respect, as the great number of Arabic population who learn Turkish as a second language (SL) is taken into consideration, it is needed to conduct research on suprasegmental features that have an important role to convey meaning within the reading and speaking a language context. To this end, the goal of this study is to investigate Arabic students’ readings who learn Turkish as a SL focusing specifically on stress and pause with the framework of suprasegmental phonology (prosody). Finch (2000) defines suprasegmentals as “units above the segmental level of phonemes” (p. 39). Hudson (2000) states that suprasegmental qualities can be seen on over more than one phone. Özbayand Çetin(2011) also statethat suprasegmental features are linguistic sub-systems which have several functions such *Corresponding author. Tel.: +90-342-360-1200/3759 E-mail address: [email protected] 216 Serdar Derman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 aslength, stress, intonation, pitch, pause and these terms are directly related to syllable, word and sentence. 1.1. Stress and Pause Ladefoged (2006) explains stressas“a kind of suprasegmental feature of utterances”, and states that it cannot be found in individual vowels or consonants but in the whole syllables. If a syllable or a word is pronounced with higher pitch than other syllables or words, it can be said that it is stressed. Accordingly, the listeners can hear that stressed syllable in a word louder, stronger, and slightly higher than the rest or the unstressed ones. To understand stress, one should consider both word and sentence stress. For the word stress, we shall distinguish two degrees of stress as shown below: primary stress is on 1 and secondary stress is on 2. categorical/k æ t ə g ʋ rk l/ 21 Collins and Mees (2013, p.130) In English, four phonetic variables appear to be the most significant indicators of stress: intensity, pitch variation, vowel quality and vowel duration. Table 1. Characteristics of stressed and unstressed syllables Stressed Unstressed 1 Intensity Articulation with greater Less breath/muscular effort breath/muscular effort Perceived as having less Perceived as greater loudness loudness 2 Pitch Marked change in pitch Syllables tend to follow the pitch trend set by previous stressed syllable 3 Vowel quality May contain any vowel Generally have central vowels (except /e/) /e x k/ or syllabic consonants Vowels have clear (peripheral) Vowels may have centralized quality quality Diphthongs have clearly Diphthongs tend to have a defined glide much reduced glide 4 Vowel Vowels have full length Vowels are considerably duration shorter Collins and Mees (2013, p.130) Collins and Mees (2013) explain that many of the potential stress of words are lost in connected speech. So, in general, words which carry little information normally lose stress. These are the words important for the structure of the sentence, i.e. function words (articles, auxiliary verbs, verb be, prepositions, pronouns, conjunctions). Content words (nouns, main verbs, adjectives, most adverbs), which carry high information load, are normally stressed. “I’ve heard that Jack and Jane spent their holidays in Jamaica. F F C F C F C C F C F C (C = content word, F = function word)” (p.135) . SerdarDerman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 217 Pause is one of the prosodic features which accomplish the phonological and delimiting function in sentence level and its constituents. At the end of 60s, the pioneering figure of pausology Goldman- Eisler (1968) created pausological measurement conventions which can be done through detailed analysis of spectrographic printouts. Following them, various linguists have studied pause from different point of views. Vogel (1986) carried out a study on hesitation pauses. For him, pauses are used to show the limits of prosodic components such as phonological utterance and intonational phrases-elements which incorporate phonological, syntactical, semantic, and pragmatic messages that can be found in clauses and its parts; furthermore, they can be found in phrases or sentences. In this way, one can find whether there is hesitation through the predictability of relation between pause points and prosodic components. Later, Taylor and Black (1998) developed a formula to calculate students’ correct pauses, which will be explained in data analysis. From a descriptive point of view, pauses are divided into two categories in general use. The first one is physical and linguistic pauses which contain intra-segmental and inter-lexical pauses, the latter are psychological and psycholinguistic which include pauses in terms of their origin and function. As seen above, pauses have an important role in determining the speaker as fluent or not (Zellner, 1994). When it comes to the relationship of stress and pause in reading aloud and speaking skills, Schwanenflugel et al (2004) claim that one should consider (a) perceived changes in pitch, (b) stress or loudness, (c) duration and pausing, because prosodic reading allow readers to chunk group of words into phrases and meaningful units in terms of syntactic structure of text . For them, age is an important factor to use both stress and pausing strategies appropriately. Namely, to read prosodically, age is a significant factor to understand and use prosodic features in spoken language, for instance; 8 year old children cannot perform prosodic stress patterns to convey meaning and to comprehend the difference between these sample sentences: Beth is already at the party and Beth is already at the party. Schwanenflugel et al (2004) also state that pauses may be effective through commas like Lesley came, she saw, and she conquered, but may not be effective for these sentences: Lesley wanted the one with the red, white, and blue sprinkles. Through these examples, it can be said that as written texts have longer sentences which tax short term memory, prosodic features must be abstracted by oral readers while reading aloud. 1.2. Research questions This study tries to describe TSL learners’ read speech and reveal their use of suprasegmental features, particularly the use of stress and pause. To this aim this study seeks to find answers to the following research questions: 1. To what extent are Arabic TSL learners successful in using primary word stress while reading aloud? 2. To what extent are Arabic TSL learners successful in using pause while reading aloud? 2. Method Since the present study aims to describe TSL learners’ use of stress and pause in their readspeech, a descriptive survey model was adopted. Karasar (1998, p. 77) defines descriptive survey model saying “current situation is tried to be described as it is in a descriptive survey model”. 218 Serdar Derman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 2.1. Sample / Participants The population of this study included TSL learners whose mother tongue is Arabic. The study group consisted of 34 (15 female, 19 male) Syrian students learning Turkish as a second language at a state university. The participants enrolled in Turkish courses at B2 level at TÖMER (Turkish Teaching Application and Research Center) during the 2014-2015 academic year. The learners voluntarily participated in the study and their age range was between 19 and 24. All the participants came to Turkey at the same time and they passed through each level (A1, A2, B1 and B2) in two different classes with same instructors. 2.2. Data collection procedures The data for the study were collected by having students read a text aloud. The text was chosen from a book prepared according to European language portfolio B2 level by Gazi University TÖMER (Kurt &Temur, 2013, p. 15). Students’ voice recordings were used to determine their use of stress and pausing patterns while reading aloud. While selecting the appropriate text, three experts in teaching Turkish to foreigners were asked about the relevance of the text for the study. The experts reached a consensus that the text was appropriate for their levels. The text was comprised of six sentences. Before the application, to make them familiar with the text students were given some rehearsal time. Then they were asked to read the text loud and their voices were recorded in a sound-treated room. 2.3. Data analysis Students’ voices, recorded by a digital voice recorder with a noise cut function, were analyzed using Praat and Cool Edit pro speech analysis software programs. Praat5.4.01 (2014) software was utilized to detect students’ word stress, and then the stressed syllables were determined according to the data gathered by this software. The students’ word stress placements were compared to the ones identified before by three qualified subject matter experts. Thus, the ratio of the students’ correct word stress production was determined. Students’ voice recordings were analyzed by Cool Edit Pro software, a program which could also make various sound analyses, and the pauses done by students during reading were spotted. The places of necessary pauses during loud reading were ascertained by three experts, and students’ pauses were determined accordingly. All the voice recordings were not included in the data analysis process, the ones related to the research questions were analyzed. The analyses about pausing were limited to the following issues: 1. The number of pauses in the right place. 2. The number of missing pauses although it is necessary while reading. 3. The number of pauses in the wrong place or pause insertion. 4. Total number of incorrect pauses. In the present study, the formula developed by Taylor and Black (1998) was used to calculate the ratio of the students’ pauses. The formula is as follows: 1. The number of correct pauses = (B– I)/B x 100% 2. The number of missing pauses = D /T x 100% 3. The number of pause insertion = I/(N-T)x 100% 4. Total number of incorrect pauses = I /B x 100% B= (students’ total number of pauses) . SerdarDerman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 219 I= (the number of pause insertion) D= (the number of missing pauses) T= (the number of necessary pause) N= (the number of total pauses that could be appropriate in the text). 3. Results Table 2 below presents the correct places of primary stressed syllables and the frequencies of the students’ stress placements. Table 2. The frequencies of students’ correct and incorrect stress placements for the first sentence of the text. Word Syllable1 Syllable2 Syllable3 Syllable4 Syllable5 Syllable6 insanla'rın 11,76 35,29 5,88 47,06 yaşamla'rı 11,76 41,18 20,59 26,47 süre'since 2,94 26,47 14,71 55,88 ihti'yaç 8,82 11,76 76,47 duydukla'rı 11,76 41,18 0,00 47, 06 'belki 17,65 82,35 sami'mi 5,88 47,06 47, 06 dosta'ne 11,76 41,18 47,06 ilişki'ler 2,94 14,71 58,82 23, 53 'iken 17,65 82,35 gi'derek 8,82 38,24 52, 94 menfaa'te 5,88 44,12 0,00 50, 00 dö'nük 8,82 91,18 ilişki'ler 5,88 20,59 64, 71 8,8 2 için'de 8,82 23,53 67,65 olmala'rı 14,71 26,47 5,88 52, 94 bire'yi 17,65 14,71 67,65 top'lum 35,29 64,71 için'de 8,82 17,65 73, 53 yalnızlı'ğa 11,76 23,53 2,94 61, 76 sürüklemek'tedir 11,76 44,12 8,82 11,76 5,8 8 17, 65 When the table is examined it is seen that learners put the primary stress correctly for the words of the first sentence respectively as “insanların” 47.06%, “yaşamları”26.47%, “süresince” 14.71%, “ihtiyaç” 76.47%, “duydukları” 47.06%, “belki” 17.65%, “samimi” 47.06%, “dostane” 47.06%, “ilişkiler” 23.53%, “iken” 17.65%, “giderek” 38.24%, “menfaate” 50%, “dönük” 91.18 %, “ilişkiler” 8.82%, “içinde” 67.65%, “olmaları” 52.94%, “bireyi” 67.65%, “toplum” 67.71, “içinde” 73.53%, “yalnızlığa” 61.67%, “sürüklemektedir” 5.88%. As can be seen from Table 2, learners successfully put the primary stress on the words “dönük” (91.18%) and “içinde” (73.53% and 67.65). However, they put the primary stress incorrectly on the words “sürüklemektedir” (5.88%), “süresince” (14.71%) and “belki” (17.65%). It was observed that 220 Serdar Derman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 students’ correct stress ratio was low in words whose primary stress is not on the final syllable such as belki, iken, sürüklemektedir, süresinceand so on. It can be said that because the word stress is generally put on the final syllable in Turkish, learners’ overgeneralizations lead them to make mistakes while reading. The words “ilişkiler” and “içinde” took place twice in the sentence; however, the correct stress ratio for the first time for “ilişkiler” was 23.53%, it was 8.82% for the second time; the correct stress ratio for the first time for “içinde” was 67.65%, it was 73.53% for the second time. Many of the learners put stress on –ki syllable in “ilişkiler” (58.82% in the first and 64.71% in the second time it occurs). This can be explained by the nature of /k/ sound in –ki syllable because it is a voiceless plosive stop. The changes of the stress ratio for the word “içinde” can be accepted as normal when the dynamics of read speech are considered. Table 3 shows the correct places of primary stressed syllables and the frequencies of the students’ stress placements for the second sentence of the text. Table 3. The frequencies of students’ correct and incorrect stress placements for the second sentence of the text Word Syllable1 Syllable2 Syllable3 Syllable4 Syllable5 Çıkarla'rın 11,76 38,24 5,88 44,12 plan'da 2,94 47,06 50,00 oldu'ğu 20,59 32,35 47,06 toplum'da 29,41 29,41 41,18 ilişki'ler 2,94 23,53 38,24 35,29 sev'gi 11,76 88,24 say'gı 2,94 97,06 yardımlaş'ma 2,94 14,71 14,71 67,65 dayanış'ma 5,88 5,88 38,24 50,00 'özveri 29,41 26,47 44,12 gi'bi 17,65 82,35 değerle're 14,71 20,59 8,82 55,88 de'ğil 52,94 47,06 konu'su 35,29 32,35 32,35 çıkarla'ra 14,71 29,41 2,94 52,94 daya'lı 17,65 0,00 82,35 o'larak 11,76 29,41 58,82 yürümek'tedir 0,00 20,59 55,88 8,82 14,71 When we examine the table we can see that learners put the primary stress correctly in the ratio as “çıkarların” 44.12%, “planda” 50.00%, “olduğu” 47.06%; “toplumda” 41.18%, “ilişkiler” 35.29%; “sevgi” 88.24%, “saygı” 97.06%, “yardımlaşma” 67.65%, “dayanışma” 50.00%, “özveri” 29.41%, “gibi” 82.35%, “değerlere” 55.88%, “değil” 47.06%, “konusu” 32.35%, “çıkarlara” 52.94%, “dayalı” 82.35%, “olarak” 29.41%, “yürümektedir” 8.82% . When the above table is examined, it is seen that students successfully put the stress on the correct syllable in words “saygı” (97.06%), “sevgi” (88.24%) and “dayalı” (82.35%). These words have two or three open syllables and they have the primary stress on the final syllable as it is generally the case in Turkish. However, students made mistakes in words like “yürümektedir” (8.82%), “özveri” (29.41%) ve“olarak” (29.41%); their success rate was very low in these words. The reason for this . SerdarDerman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 221 could be that these words have three or more syllables with stress on a different syllable other than the final one. Table 4 presents the correct places of primary stressed syllables and the frequencies of the students’ stress placements. Table 4. The frequencies of students’ correct and incorrect stress placements for the third sentence of the text Word Syllable1 Syllable2 Syllable3 Syllable4 Syllable5 Syllable6 Syllable 7 Bura'dan 0,00 41,18 58,82 konu'yu 20,59 26,47 52,94 şu'na 38,24 61,76 bağlayabi'lirim 5,88 41,18 2,94 26,47 2,94 20,59 beklentile'rin 23,53 14,71 29,41 0,00 35,29 menfaatle'rin 17,65 35,29 8,82 5,88 29,41 hâ'kim 11,76 88,24 oldu'ğu 5,88 23,53 70,59 ilişki'ler 8,82 20,59 44,12 26,47 dünyasın'da 23,53 35,29 14,71 26,47 in'san 26,47 73,53 çevresindekile'ri 11,76 14,71 20,59 5,88 8,82 0,00 38,24 hu'zur 5,88 94,12 başarısı'nın 5,88 11,76 23,53 0,00 58,82 önün'de 2,94 11,76 85,29 teh'dit 5,88 94,12 o'larak 0,00 17,65 82,35 algılayabi'lir 14,71 44,12 8,82 14,71 8,82 8,82 When the table is examined it is seen that learners put the primary stress correctly for the words which formed the third sentence respectively as “buradan” 58.82%, “konuyu” 52.94%, “şuna” 61.76%, “bağlayabilirim” 2.94%, “beklentilerin” 35.29%, “menfaatlerin” 29.41%, “hâkim” 88.24%, “olduğu” 70.59%, “ilişkiler” 26.47%, “dünyasında” 26.47%, “insan” 73.53%, “çevresindekileri” 38.24%, “huzur” 94.12%, “başarısının” 58.82%, “önünde” 85.29%, “tehdit” 94.12%, “olarak” 17.65%, “algılayabilir” 8.82%. For this sentence, learners successfully put the primary stress on the correct syllable in words “huzur” and “tehdit” (94.12%) and “önünde” (85.29%). This success might stem from their native language because these words, specifically “huzur” and “tehdit”, were derived from Arabic. On the other hand, the learners put the primary stress on a wrong syllable in words “bağlayabilirim” (2.94%), “algılayabilir” (8.82%), “olarak” (% 17.65). Like in the previous sentences, these words have three or more syllables with stress on different syllable other than the final one. In other words, the longer the word, the more likely learners are to put the primary stress on the wrong place. Table 5 illustrates the correct places of primary stressed syllables and the frequencies of the students’ stress placements for the fourth sentence of the text. 222 Serdar Derman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 Table 5. The frequencies of students’ correct and incorrect stress placements for the fourth sentence of the text Word Syllable1 Syllable2 Syllable3 Syllable4 Syllable5 Syllable6 'Böylelikle 14,71 38,24 20,59 26,47 ilişkilerin'den 2,94 8,82 23,53 0,00 5,8 8 58, 82 endi'şe 23,53 35,29 41,18 du'yarak 23,53 11,76 64,71 insanlar'dan 5,88 14,71 11,76 67, 65 uzaklaşabi'lir 5,88 32,35 17,65 29,41 0,0 0 14, 71 It is clear from the table that learners’ correct placement of the primary stress frequencies are 14.71% for “böylelikle”, 58.82% for “ilişkilerinden”, 41.18% for “endişe”, 11.76% for “duyarak”, 67.65% for “insanlardan”, 14.71% for “uzaklaşabilir”. The most successfully stressed words for the fourth sentence were “insanlardan”(67.75%) and “ilişkilerinden”(58.82%). Although these words have more than three syllables, learners placed the correct stress on the right syllable. This can be explained by the nature of –den/–dan ablative suffixes since they are among stressed affixes. However, similar to the previous sentences, we can see that learners again had the most mistaken stress placement on the words with three or more syllables like “duyarak” (11.76%) and “uzaklaşabilir” (14.71%). Table 6 demonstrates the correct places of primary stressed syllables and the frequencies of the students’ stress placements. Table 6. The frequencies of students’ correct and incorrect stress placements for the fifth sentence of the text Word Syllable1 Syllable2 Syllable3 Syllable4 Syllable5 Syllable6 Yalnız'lık 11,76 20,59 67,65 fer'din 23,53 76,47 di'ğer 26,47 73,53 insan'larla 11,76 26,47 2,94 58,82 münasebe'ti 2,94 11,76 14,71 20,59 50,00 sırasın'da 11,76 20,59 8,82 58,82 yaşa'nan 11,76 17,65 70,59 problemler'den 0,00 26,47 17,65 5,88 50,00 kaynaklanmak'tadır 17,65 38,24 14,71 8,82 8,82 11,76 When it is examined, it is seen that learners put the primary stress correctly for the words which formed the fifth sentence as “yalnızlık” (67.65%), “ferdin” (76.47%), “diğer” 73.53%, “insanlarla” (2.94%), “münasebeti” (50.00%), “sırasında” (58.82%), “yaşanan” (70.59%), “problemlerden” (50.00%), “kaynaklanmaktadır” (8.82%). It was observed that learners pronounced the words “ferdin” (76.47%), “diğer” (73.53%), and “yalnızlık” (67.65%) by placing the correct stress. On the other hand, they put the stress on the wrong syllables in the words “insanlarla” (2.94%) and “kaynaklanmaktadır” (8.82%). Along with the words in the four previous sentences, the correct stress rates are high in two or three syllable words having . SerdarDerman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 223 primary stress on the final syllable. Nevertheless, the correct stress rates are too low in three or more syllable words having primary stress on syllables other than the final. Table 7 shows the correct places of primary stressed syllables and the frequencies of the students’ stress placements for the last sentence of the text. Table 7. The frequencies of students’ correct and incorrect stress placements for the sixth sentence of the text Word Syllable1 Syllable2 Syllable3 Syllable4 Syllable5 Syllable6 'ise 61,76 38,24 ço'ğu 23,53 76,47 za'man 23,53 76,47 anlaşı'lamamak 5,88 0,00 8,82 17,65 14,71 52,94 anla'yamamaktan 8,82 8,82 14,71 2,94 5,88 58,82 ile'ri 5,88 20,59 73,53 gelmek'tedir 32,35 47,06 20,59 0,00 When we examine the table we can see that learners put the primary stress correctly in the ratio asise” (61.67%), “çoğu” (76.47%), “zaman” (76.47%), “anlaşılamamak” 17.65%, “anlayamamaktan” (14.71%), “ileri” (73.53%) and “gelmektedir” (20.59%). As is clear from the table, learners put the primary stress more successfully in the words “çoğu” (76.47%), “zaman” (76.47%) and “ise” (61.76%). They were least successful with the words “anlaşılamamak” (17.65%) and “anlayamamak” (14.71%) in this sentence, because these words are the noun form of negative compound verbs, and this complex structure may lead learners to place the stress on the wrong syllable. Table 8 below presents the frequency of average correct stress for each sentence. Table 8. The frequency of average correct stress per sentence Sentence number Average correct stress (%) 1 45.10 2 52.29 3 54.90 4 34.80 5 50.98 6 48.74 Total average correct stress 47.80 As can be seen in Table 8, the general average correct stress placement score of the learners is 47.80%, which can be said to be low according to their proficiency levels. The reason for this could be the lack of training about the relation between the stress and meaning of a word. In addition to this, the framework for teaching Turkish as a foreign/second language does not contain a detailed syllabus for teaching suprasegmental features. Table 9 shows the correct pause rates for each student. 224 Serdar Derman et al. / Journal of Language and Linguistic Studies, 13(1) (2017) 215-231 Table 9. Correct pause rates for each student Student Correct pause rate Student Correct pause number (%) number rate (%) 1 16.67 18 53.33 2 60.87 19 41.46 3 50.00 20 36.36 4 44.00 21 35.00 5 56.00 22 29.79 6 44.00 23 28.57 7 50.00 24 32.56 8 38.46 25 31.11 9 32.65 26 41.18 10 65.38 27 31.25 11 42.86 28 32.00 12 31.58 29 35.00 13 51.85 30 38.64 14 45.16 31 38.46 15 37.14 32 32.65 16 35.14 33 32.35 17 34 38.24 Total average 39.69 As it is shown, the high est correct pause rate was fo und to be 6 5.38% and the lowest correct pause was 16.67%, and total average rate was 39.69%. Table 10 presents the missing pause rates for each student. Table 10. Missing pause rates for each learner Student Missing pause rate Student Missing pause rate number (%) number (%) 1 83,33 18 22,22 2 22,22 19 5,56 3 55,56 20 11,11 4 38,89 21 22,22 5 22,22 22 22,22 6 38,89 23 22,22 7 16,67 24 22,22 8 16,67 25 22,22 9 11,11 26 22,22 10 5,56 27 16,67 11 16,67 28 11,11 12 0,00 29 22,22 13 22,22 30 5,56 14 22,22 31 16,67