The Pronunciation of English Voiceless Plosives /p/, /t/, and /k/ by Palestinian EFL Learners.

(1)

Master Thesis

The Pronunciation of English Voiceless Plosives

/p/, /t/, and /k/ by Palestinian EFL Learners.

by

Montaser Motia Ujvári

S3227197

MA Multilingualism

Department of Applied Linguistics and Frisian Language and

Culture

Faculty of Arts

University of Groningen

First Supervisor

Dr. Charlotte Gooskens

Second Supervisor

Dr. Dicky Gilbers

(2)

List of Tables and Figures

Figure 1. The VOT category of English and Arabic plosives. Figure 2. The effect of vowels’ length on VOT.

Figure 3. The effect of stress and words’ context on VOT.

Figure 4. The effect of English phonetic instruction on pronunciation. Figure 5. A visual illustration of the thesis hypotheses.

Figure 6. The outliers in the study. Figure 7. VOT of Arabic voiced /b/.

Figure 8. The interaction between groups in the pronunciation of the Arabic /b/. Figure 9. VOT of Arabic voiceless /t/.

Figure 10. The interaction between groups in the pronunciation of the Arabic /t/. Figure 11. VOT of Arabic voiceless /k/.

Figure 12. The interaction between groups in the pronunciation of the Arabic /k/. Figure 13. VOT of the English voiceless plosive /p/ in authentic pronunciation.

Figure 14. The interaction between groups in the authentic pronunciation of the English /p/. Figure 15. VOT of the English voiceless plosive /t/ in authentic pronunciation.

Figure 16. The interaction between groups in the authentic pronunciation of the English /t/. Figure 17. VOT of the English voiceless plosive /k/ in authentic pronunciation.

(4)

Figure 19. VOT of the English voiceless plosives /p/, /t/, /k/ by male English native speakers. Figure 20. VOT of English voiceless plosive /p/ in converged pronunciation.

Figure 21. The interaction between groups in the converged pronunciation of the English /p/. Figure 22. The difference between groups in the convergence to a male English native

speaker’s pronunciation of the English /p/.

Figure 23. The interaction between groups in the convergence to a male English native speaker’s pronunciation of the English /p/.

Figure 24. VOT of English voiceless plosive /t/ in converged pronunciation.

Figure 25. The interaction between groups in the converged pronunciation of the English /t/. Figure 26. The difference between groups in the convergence to a male English native speaker’s pronunciation of the English /t/.

Figure 27. The interaction between groups in the convergence to a male English native speaker’s pronunciation of the English /t/.

Figure 28. VOT of English voiceless plosive /k/ in converged pronunciation.

Figure 29. The interaction between groups in the converged pronunciation of the English /k/. Figure 30. The difference between groups in the convergence to a male English native

speaker’s pronunciation of the English /k/.

Figure 31. The interaction between groups in the convergence to a male English native speaker’s pronunciation of the English /k/.

Figure 32. A visual illustration of the results from the three experiments. Figure 33. The proposed VOT category of Palestinian Arabic Plosives

Table 1. VOT of English plosives in two studies. Table2. VOT of Arabic plosives in two studies. Table 3. The hypotheses of the thesis.

Table 4. The list of Arabic words used in the First experiment. Table 5. The list of English words used in the second experiment. Table 6. The lost of English words used in the third experiment. Table. 7 A summary of the statistical results of Arabic pronunciation.

(5)

Acknowledgments

First of all, I want to thank my thesis supervisor Dr. Charlotte Gooskens who gave me a constant feedback on my work. It has been an honor to be under her supervision. Also, I want to thank everyone whom I took courses with during my study, namely Prof. Goffe Jensma, Dr. Nana Hilton, Dr. Hanneke Leorts, Dr. Joana Duarte, Dr. Eva Daussà, and Dr. Jelena Golubovic. Their courses were very informative and enjoyable. I also want to thank Dr. Dicky Gilbers, my second thesis supervisor, who gave me helpful feedback on my thesis proposal. Although we have met only once, it has been a pleasure to meet him, too. They are all great people with great personalities and will not hesitate to give a helping hand if needed.

(6)

Abstract

A common English pronunciation error among Arabs EFL learners is the pronunciation of the voiceless plosive /p/ as a voiced plosive [b]. A main marker that distinguishes between voiceless plosives and voiced plosives is the VOT (Voice Onset Time). This study aims at investigating the pronunciation of the English voiceless plosives /p/, /t/, and /k/ by Palestinian EFL learners and the effect of gender and English phonetic instruction on VOT duration of these plosives. Therefore, three experiments were carried out.

The first experiment aims at investigating VOT of the Palestinian Arabic plosives /b/, /t/ and /k/. It was found that Palestinian EFL learners produced lead voicing VOT for the Arabic voiced plosive /b/, whereas the participants’ VOT duration for the Arabic voiceless plosives /t/ and /k/ fell in between short lag VOT category and long lag VOT category. A two-way ANOVA test showed that females produced significantly shorter lead voicing VOT duration for the voiced /b/ and longer VOT duration for the voiceless /k/ than males. However, there was no effect of gender on the VOT duration of /t/. Furthermore, there was no effect of English phonetic instruction on VOT of Palestinian Arabic plosives. No interaction between the groups was found. The second experiment aims at investigating the effect of gender and English phonetic instruction on VOT of the English voiceless plosive /p/, /t/, and /k/ in the participants’ authentic English pronunciation, and to see if there is an interference from L1. Authentic pronunciation refers to the participants’ natural pronunciation of English. A two-way ANOVA test showed that the effect of gender on VOT of English voiceless plosives in authentic English pronunciation was not significant. On the other hand, it was found that the effect of English phonetic instruction on VOT duration of English voiceless plosives was significant in the case of /p/ only. No interaction between the groups was found.

(7)

convergence to a male English native speaker's pronunciation of the voiceless plosives /p/, /t/, and /k/ was not significant. No interaction between the groups was found.

(8)

1. Introduction

When I was still doing my Bachelor degree in English Language and Literature, my English Phonetics teacher told the class a joke: "An Arabic speaker went to the USA and wanted to park his car in front of the shopping center. He saw a police officer in the parking lot and asked him "can I bark here?" The officer replied "you can bark anywhere you want" ". All of the students giggled, and they immediately knew what the joke was about. However, to some people, especially those who are not familiar with the Arabic language, it may not make any sense, and one may wonder why the man said bark instead of park. In fact, this joke points to a common pronunciation error committed by many native Arabic speakers when speaking English: pronouncing the /p/ sound as [b]. Although the MSA (Modern Standard Arabic) has eight oral plosives (see section 2.2.2), which Carr (1999) defines as sounds which are produced with a complete closure in the mouth, the voiceless plosive /p/, which exists in many languages, is absent in Arabic (Ryding, 2014). As a result, some Arab EFL learners pronounce it as [b] due to L1 interference.

Some Arab EFL learners are actually prone to making such an error; their pronunciation may be misleading in some words where /p/ and /b/ are minimal pairs. One might say something that does not make sense at all like "I ate three bears this morning". Sometimes, things may become awkward; telling someone who may have some cultural or religious inquiries concerning Arab communities that "we bray five times a day" seems quite shocking at first glance. Therefore, it becomes important to distinguish the differences between Arabic sounds and English sounds in order to avoid any possible embarrassment and misunderstanding. The reason I am doing this acoustic study is that there is little research done concerning the Arabic plosives. In addition, there has been little research concerning Palestinian EFL learners’ pronunciation of English. Hence, I want to conduct a research that is original and relevant to my native language and local dialect.

(9)

(natural) English pronunciation, and /p/, /t/, and /k/ in their English pronunciation after hearing a native speaker (converged English pronunciation).

After this introduction, the "voice onset time" concept will be explained. After that, relevant literature concerning the voice onset time of English and Arabic plosives will be introduced. In addition, the factors which may have an effect on voice onset time will be demonstrated. Following the literature review section, the research questions and sub-questions will be addressed, and the hypotheses will be introduced. After that, the research methodology will be explained; the independent and dependent variables will be introduced and the procedures will be explained step by step. Following the methodology section, the results will be reported statistically, and the proposed hypotheses will be confirmed or refuted. Then, in the discussion section, the results will be interpreted and linked to the research questions and to the findings in the literature review section. Then, in the conclusion section, the findings of the study will be summed up, recommendation for future research concerning this topic will be given, the limitations of the study will be mentioned, and practical applications will be suggested. Finally, the appendices will be attached at the end of the research right after the references, which consist of the materials that were used in this study and the results of each individual participant.

2. Literature Review.

2.1 Arab EFL Learners' Pronunciation of The Sound /p/

Many Arab EFL learners have a pronunciation error in common: pronouncing the sound /p/ as [b]. In a study conducted to investigate English pronunciation problems of Sudanese EFL learners at Sudan University of Science and Technology, Hassan (2014) found that only fourteen participants (28%) out of fifty pronounced the sound /p/ correctly. Elmahdi & Khan (2015) found similar results in a study which investigated English pronunciation problems of sixty Saudi secondary school students. They found that more than 89% of the participants had a problem in pronouncing the sound /p/ in word initial position, 87% in word medial position, and 81% in word final position due to L1 interference. The results of both studies depended on the researchers' own judgments of the participants' pronunciation through direct observation inside classrooms and tape recordings of the participants reading English sentences.

2.2 Voice Onset Time

(10)

word-initial position fall into three categories: negative VOT (lead voicing or prevoicing), short lag VOT, and long lag VOT. Negative VOT occurs when voicing starts before the release of the plosive (approximately -30 ms or more), short lag VOT occurs when voicing starts shortly after the release of the plosive (approximately 0 to +30 ms), and long lag VOT occurs when voicing starts well after the release of the plosive (+50 ms or more). (See also Arabski & Wojtaszek 2011).

2.2.1 Research on VOT of English Oral Plosives

English utilizes three places of articulation for the production of oral plosives: bilabial, alveolar, and velar. In each of these places of articulation one voiced and one voiceless plosive are produced; voiceless /p/ and voiced /b/ in the bilabial position, voiceless /t/ and voiced /d/ in alveolar position, and voiced /g/ and voiceless /k/ in velar position (Ogden 2009). Each plosive has different VOT duration with a clear boundary between voiced and voiceless plosives.

In a study investigating the VOT of oral plosives among four American English native speakers, Lisker and Abramson (1964) found that three of the participants produced short lag VOT for voiced plosives and long lag VOT for voiceless plosives in pronouncing single words. Klatt (1975) found similar results in his study on three monolingual English native speakers. In spoken words in the context of the sentence "Say ____ instead", he found that English voiced plosives belong to the short lag VOT category, whereas English voiceless plosives belong to the long lag VOT category with a clear boundary between voiced and voiceless plosives.

Table 1. The voice onset time of English oral plosives in two different studies.

Research on English VOT

Bilabial Alveolar Velar /b/ /p/ /d/ /t/ /g/ /k/

Lisker and Abramson (1964:394)

N=3

Stimuli: word reading

+1 ms +58 ms +5 ms +70 ms +21 ms +80 ms Klatt (1975:689) N=3 Stimuli: word in "Say___ instead" +11 ms +47 ms +17 ms +65 ms +27 ms +70 ms

2.2.2 Research on VOT of Arabic Oral Plosives

(11)

(voiced alveolars), /k/ (velar), /q/ (uvular), and /ʔ/ (glottal). However, there is little research which looks into the voice onset time of these sounds. In a study investigating Lebanese monolingual and bilingual children's pronunciation, Khattab (2002) found that Arabic monolinguals produced negative VOT for voiced plosives and short lag VOT for voiceless plosives. In her study, the VOT results of the three monolingual children were reported separately since the aim was to point out the effect of age on VOT. Table 2 shows the results of the three participants after calculating the mean VOT duration for the participants. Khattab (2002) explains that the /p/ and /g/ are not part of the sound inventory of Lebanese Arabic, but they may occur in loan words. Therefore, the /p/ and /g/ are presented in figure 1 as part of the Arabic plosives as well.

Figure 1 shows places of plosives they occupy in the VOT continuum in English and Arabic (Khattab 2002: 96).

On the other hand, in a study comparing the VOT of CA (Classical Arabic) and MSA (Modern Slandered Arabic), Aldahri (2012) found that voiced plosives belong to the short lag VOT category, whereas voiceless plosives belong to the long lag VOT category in both varieties.

Table 2. The voice onset time of Arabic plosives in two different studies.

Research on Arabic VOT Bilabial Alveolar Velar

/b/ /t/ /d/ /k/

Aldahri (2012:693) N=3

Stimuli: word reading (MSA) +14 ms +50 ms +15 ms +53 ms Khattab (2012:69)

N= 3

Stimuli: word elicitation (Lebanese Arabic)

-45 ms +22 ms -44 ms

(12)

2.3. Factors That Affect VOT

2.3.1 The Relation Between VOT and Gender

There are conflicting results on the relation between gender and VOT. In a study which involved twenty monolingual speakers of American English (ten females and ten males), Robb, Gilbert, and Lerman (2005) found that females have significantly longer VOT for the voiceless plosives compared to males in non-laboratory setting due to biological differences between females and males “Presumably, the gender differences found for VOT production reflect fundamental biological differences in the various speech subsystems” (p.126). Whitesand and Marshal (2001) found similar results in a study which investigated the pronunciation of thirty English monolingual children from different age groups. They point out that females produced longer VOT duration than males.

On the other hand, Morris, McCrea, and Herring (2008) point out that there is no relation between VOT and the sexes. In a study which involved eighty English monolingual participants (forty females and forty males), they found that there is no significant VOT difference between females and males in the production of voiced and voiceless plosives although females tend to have slightly longer VOT for the voiceless plosives.

2.3.2. Linguistic Factors That Affect VOT

There are some linguistic factors that may affect the voice onset time of oral plosives. One of the phonological factors is the duration of the vowel following the plosive. In a study investigating the pronunciation of eight English native speakers in single words reading, Port and Rotunno (1979) found that there is a significant positive correlation between the VOT duration of the plosive in word-initial position and the duration of the following vowel. In other words, the VOT duration of the plosive is longer when followed by a long vowel than when followed by short a vowel.

(13)

Another phonological factor which may affect the length of VOT of voiceless plosives is attributed to the location of stress. Lisker and Abramson (1967) believe that stress is significant for VOT "since we have reason to believe that the feature of aspiration is directly related to the timing of voice onset, it follows that we should expect differences in stress to be reflected in our VOT measurements" (p.15). In a study which involved three native speakers of American English, they found that voiceless plosives have longer VOT duration when they are in a stressed syllable than when they are in an unstressed syllable although the difference is not significant.

In addition, Lisker and Abramson (1967) point out that the context of the word is of a great significance. They point out that words uttered in isolation have significantly longer VOT than when they are uttered in connected speech "Where the difference between mean values for /ptk/ in isolated words as against sentences is about 25 msec, the comparable difference between stressed and unstressed /ptk/ is only about 6 msec" (Lisker and Abramson 1967: 16)

Figure 3 shows the mean VOT duration for the voiceless plosives /p/, /t/, and /k/ in syllable initial position in isolated words and connected speech (Lisker & Abramson 1967: 17).

2.3.3 The Effect of Explicit Phonetic Instruction on Pronunciation

(14)

(2009) found that the EFL learners trained to discriminate between /p/ and /b/ sounds showed an improvement in their production of the plosive consonant contrasts.

Figure 4 shows that the difference in VOT duration for the sound /p/ before (T1) and after (T2) receiving phonetic training compared to native speakers (Aligia-Garcia & Mora 2009: 19).

2.4 The Relation Between Gender and Phonetic Convergence

There are conflicting results on the relation between gender and phonetic convergence. Namy, Nygaard, and Sauerteig (2002) point out that people often engage in vocal accommodation in which they adapt to the characteristics of their conversational partners’ accent. In a shadowing task in which the participants were asked to repeat isolated words spoken by various talkers without being explicitly asked to engage in accommodation, they found that female participants showed more phonetic convergence to the utterances they heard in L2 than males. In addition, they found that all participants accommodated to male speakers more than female speakers. However, Pardo’s (2006) findings suggest the opposite. In a map task in which participants were divided into pairs and had to communicate so that the path on the first map can be duplicated on the second map, he found that male speakers showed more phonetic convergence than female speakers. He also found that the effects of the talker's sex did not follow the predictions suggested by Namy, Nygaard, and Sauerteig (2002).

3. Research Questions and Hypotheses

3.1 Research Questions and Sub Questions

(15)

VOT category and voiceless plosives to the long lag VOT category. The question that arises here is to which VOT categories do the Palestinian Arabic voiced plosive /b/ and voiceless plosives /t/ and /k/ belong?

Regarding the VOT of English plosives, the literature collected in the literature review section suggests that English voiced plosives belong to the short lag VOT category while voiceless plosives belong to the long lag VOT category. However, the case may not be the same for Palestinian EFL learners; there might be an interference from L1 which may affect their pronunciation of English plosives. The question that arises here is how Palestinian EFL learners pronounce the English voiceless plosives /p/, /t/, and /k/ and whether there is an interference from their L1? Do Palestinian EFL learners show a difference between the pronunciation of Arabic and English plosives?

Some studies suggest that females have longer VOT duration for the voiceless plosives than males while other studies suggest that there is no significant effect of gender on VOT. The question that arises here is which gender group has longer VOT duration for the Arabic voiceless plosives. In addition, which gender group has longer VOT duration for the English plosives in authentic and converged English pronunciation?

Furthermore, some studies suggest that females converge more to English native speakers' accent than males. However, other studies suggest the opposite and argue that males show more convergence to English native speakers' accent. The question that arises here is which gender group among Palestinian EFL learners shows more convergence to English native speakers' pronunciation of the English voiceless plosives?

(16)

3.2 Hypotheses

The different hypotheses and expectations are presented in table 3 on the basis of the previously discussed literature in the literature review section.

Table 3. The hypotheses of the study in the three experiments.

Females vs. Males English Phonetic instruction vs. no phonetic

instruction Palestinian

Arabic

pronunciation

1) Both groups will have lead voicing VOT for the voiced plosive /b/ and short lag VOT for the voiceless plosives /t/ and /k/. 2) Females will have longer VOT duration for the voiceless plosives /t/ and /k/ and shorter lead voicing duration for the voiced plosive /b/ than males.

3) Both groups will have lead voicing VOT for the voiced plosive /b/ and short lag VOT for the voiceless plosives /t/ and /k/.

4) There will be no significant difference between the groups regarding the VOT duration of the /b/, /t/, and /k/ sounds.

Authentic English

pronunciation

5) Both groups will have short lag VOT for the voiceless plosive /p/ due to L1 interference, and short lag VOT duration for the /t/ and /k/ plosives (similar to their

corresponding Arabic plosives). 6) Females will have longer VOT duration for the voiceless plosives.

7) Both groups will have short lag VOT for the voiceless plosives /p/, /t/, and /k/.

8) The group who received English phonetic instruction will have significantly longer VOT duration for the voiceless plosives /p/, /t/, and /k/ than the group who did not receive English phonetic instruction.

Converged English

pronunciation

9) Both groups will show

convergence to an English native speaker's pronunciation of the voiceless plosive/p/, but will not show convergence to the

pronunciation of /t/ and /k/

because they will perceive them as the same plosives from experiment 2.

10) Females will still have longer VOT duration for the voiceless plosives after convergence.

11) Both groups will show convergence to a native speaker's pronunciation of the voiceless plosives /p/. However, only the group who received English phonetic instruction show convergence to the voiceless plosive /t/ and /k/ assuming they will apply English aspiration rules.

(17)

Figure 5 shows a visual illustration of the hypotheses in the three experiments (the figure does not present the exact position of the sounds as it is used to show the differences between groups).

4. Methodology

The above mentioned research questions and hypotheses were tested in a quantitative study. This section will introduce the participants in the study, materials, procedure, and analyses of the data.

4.1 Subjects

(18)

= 21). The participants in the control group were selected randomly from the Faculties of Economics and Engineering, whereas the participants in the experimental group were selected from the Faculty of Arts, specifically from The English Language and Literature Department. In addition, each group was divided into two groups based on gender; 12 females and 9 males in the control group and 9 females and 12 males in the experimental group.

The participants' full VOT results are reported in Appendix 6. In this data set participants were given codes to identify which groups they belong to. For example, participant number 1 was coded like this: P1-MP, where the left part of the dash (-) refers to the participants' number, and the right part of the dash refers to which independent groups they belong to; the first letter refers to the gender (Male (M) or Female(F)), whereas the second letter indicates if they had taken English phonetics and phonology courses or not (Phonetic instruction (P) and No phonetic instruction (N)).

The extreme values (outliers) in this study were found in the authentic pronunciation of English /p/. These outliers are P8-FN, P10-MP, P11-FP, P16-MP, P20-FN, P41-MN, and P42-MN. I decided to include them in the study since they may reveal an effect of interference from L1 in addition to an effect of English phonetic instruction on the pronunciation of the English voiceless plosive /p/, and this is what this thesis investigates.

(19)

4.2 Materials

Before the experiment, the participants were handed a questionnaire in order to determine who to include and who to exclude from the study (see Appendix 3). The questionnaire is divided into three parts: personal information, education, and language background. The personal information section consists of two questions only which ask the participants to circle their gender and to fill in their age. The education section consists of four multiple choice questions that investigate on which English language skills their schooling focused, the level of their education (Bachelor or Master), the language of instruction in their program, and the number of English phonetics and phonology courses taken in addition to one question which investigates their age when they started learning English at school. The language background section consists of one multiple choice question which investigates if the participants have ever been to an English speaking country, two Likert scale questions which investigate how often the participants use the English language in their daily life for non-pedagogical purposes and how often they watch English movies, and one question which requires them to fill in their native language. For the first experiment, which aims at investigating the participants' pronunciation of the Arabic voiced plosive /b/ and voiceless plosives /t/ and /k/, five pictures were printed on cards; three pictures contained the Arabic target words and two distractors so that they would not be conscious about the pronunciation of the target sounds. The three target words are monosyllabic Arabic words which begin with the sounds /b/, /t/, and /k/ followed by the long vowel /i:/ in the nucleus position.

Table 3. The Arabic words that were used for the first experiment (see Appendix 1).

The word in Arabic

سيك زوم ن يت ليف ن يب

Its phonetic transcription

/ki:s/ /mo:z /ti:n/ /fi:l/ /bi:r/ Its meaning

in English

plastic bag

banana figs elephant well

(20)

Table 4. The English words that were used for the second experiment (see Appendix 2)

The English word

Key Shoes Tea Orange Peas

/ki:/ /ʃuːz/ /ti:/ /ɒrɪndʒ/ /pi:z/

In order to eliminate any effect of stress location on VOT, the target words in the first and the second experiment were deliberately chosen to be monosyllabic content words. In addition, the experiments are designed in a way that the participants say what item they see in one word only since it was found that plosives in word initial position have longer VOT when the words are uttered in isolation than when uttered in connected speech. Furthermore, all of the chosen monosyllabic target words have the long vowel /i:/ in the nucleus position preceded by the plosive which is to be investigated because it was found, as mentioned in the literature review section, that the length of the vowel following the plosive have a significant effect on the VOT duration of that plosive.

For the final experiment, which aims at testing the participants' pronunciation of the English voiceless plosives after hearing English native speakers, the pronunciation of three monosyllabic English words by male English native speakers were recorded from Cambridge Online Dictionary and Oxford Online Dictionary; the pronunciation of the words peak1 and keep2 (US pronunciation because the speaker in this variety is male) were recorded from Cambridge Online Dictionary, whereas the pronunciation of the word teach3 was recorded from Oxford Online Dictionary. The reason behind recording the word teach from Oxford Online Dictionary is that this particular word is pronounced by a female speaker only in Cambridge Online Dictionary, and the aim in the third experiment was to make the participants listen to a male English native speaker since it was reported in the literature review section that participants converge more to males speakers. The words in this experiment were deliberately chosen to be different from the words in the second experiment because the pronunciation in the second experiment might be copied in the third experiment.

(21)

Table 4. The English words that the participants were asked to repeat after a male English native speaker.

4.3 Procedures

The participants were seated in an empty, quiet room, and they were given the questionnaire to fill in before doing the experiment. They were told to ask me in case they did not fully understand some parts of the questionnaire. After that, they were asked to sit on a chair in front of me and put on the headphones. The headphones were connected to Philips DVT 1110 voice recorder; the microphone part of the headphones was connected directly to the voice recorder in order to eliminate any possible noise coming from outside the room during the experiment, and the earphone part was connected to a laptop so that the participants could hear the native speaker clearly in the third experiment. The laptop was placed in front of me, and the participants were not able to see what was displayed on the screen.

For the first experiment, the participants were asked to put on the headphones while keeping the microphone close to their mouths and say what they saw in the five pictures in a clear, moderate volume voice using their authentic Palestinian Arabic dialect and not the MSA. The participants were shown a picture of a target word in Arabic followed by a distractor. The five pictures were shown to all of the participants in the same order.

For the second experiment, the participants were asked to do the same as in the first experiment, but this time in English. Regarding the picture of peas, which I assumed would not be a familiar word to everyone, the word peas was written on a separate piece of paper. The participants who did not know this particular word in English (N= 9) were allowed to take a quick look at the written form of the word and pronounce it. The five pictures were shown to all participants in the same order.

For the third experiment, the participants were asked to repeat what they heard, and they were not explicitly asked to converge or imitate the speaker. The participants were asked to make a gesture in case they did not fully understand what they had heard so that I could play it again for them.

The English word peak teach keep

/pi:k/ /ti:tʃ/ /ki:p/

(22)

4.4 The Dependent and Independent Variables

This study aims at investigating the pronunciation of Palestinian Arabic plosives /b/,/t/ and /k/ and the English voiceless plosives /p/, /t/, and /k/ by Palestinian EFL learners by measuring the VOT of these plosives. In other words, the VOT duration of the plosives in the three experiments is the dependent variable in this study.

There are two independent variables by which the dependent variable is assumed to be affected: gender of the participants and English phonetic instruction. In this study, the VOT results of two groups will be compared: a control group consisting of females and males who did not receive any explicit English phonetic instruction, and an experimental group consisting of females and males who received intensive English phonetic instruction

.

4.5 Design and Analyses

The recordings were moved to the laptop as WAV format in order to be analyzed using PRAAT software, and the participants had their VOT results attached to their questionnaire answer sheet (see Appendix 4). After that, the participants that are not eligible for the experiment were excluded.

The results were analyzed using SPSS 23. Two-way ANOVA tests were carried with the results of the three experiments to investigate the effect of gender and English phonetic instruction on VOT. The reason behind using this statistical test is that there are two independent variables, each of them has two levels. The first independent variable "gender" consists of two levels: "females" and "males". The second independent variable "phonetic instruction" also consists of two levels: "phonetic instruction" and "no phonetic instruction". In addition, the two-way ANOVA tests make it possible to investigate if there is an interaction between levels of gender group and levels of phonetic instruction group. The α level that was chosen to reject the null hypothesis is 0.05.

5. Results

5.1 VOT of Palestinian Arabic Plosives

(23)

5.1.1 VOT of Palestinian Arabic /b/

In this part of the experiment, it was found that the Palestinian Arabic voiced plosive /b/ has negative VOT duration (M= -59.3, SD= 25.7). This means that on average, the participants had their vocal folds vibrating for around 59 ms before the release of the closure after pronouncing the voiced plosive /b/ in monosyllabic word initial position when it occurs before the long vowel /i:/ in the nucleus position.

To test the effect of gender and English phonetic instruction on VOT of the Palestinian Arabic voiced plosive /b/ and to see if there is an interaction between these two independent variables, a two-way ANOVA test was carried out.

The two-way ANOVA test showed there was a statistically significant effect of gender on VOT of the Palestinian Arabic /b/, (F (1, 38) = 7.96, p < 0.01). On average, females had significantly shorter lead voicing duration before the release of the closure in the pronunciation of the Arabic /b/ sound (M = -48, SD = 28.7) than males (M = -69.8, SD = 15).

However, the effect of English phonetic instruction on VOT of the Palestinian Arabic voiced plosive /b/ was not statistically significant,(F (1, 38) = 0.89, p = 0.35). The group who received English phonetic instruction had insignificantly longer lead voicing duration before the release of the closure in the pronunciation of the Palestinian Arabic voiced plosive /b/ (M = -62.8, SD = 25.7) than the participants who did not receive any English phonetic instruction (M = -55.8, SD = 25.4) (see figure 7).

Figure 7 shows the mean VOT duration of the Arabic voiced /b/ for the four groups.

There was no significant interaction between gender and English phonetic instruction in the pronunciation of /b/, (F (1, 38) = 0.62, p = 0.30) (see figure 8).

(24)

Figure 8 shows the interaction between groups in pronouncing the Palestinian Arabic voiced plosive /b/.

5.1.2 VOT of Palestinian Arabic /t/

It was found that the Palestinian Arabic voiceless plosive /t/ has positive VOT duration (M = 36.5. SD = 13.9). This means that on average, the participants had the burst of the voiceless plosive /t/ in monosyllabic word initial position around 36.5 ms earlier to the vibration of their vocal folds when this plosive occurs before the long vowel /i:/ in nucleus position.

A two-way ANOVA test showed that the effect of gender on VOT of the Palestinian Arabic voiceless plosive /t/ was not statistically significant (F (1, 38) = 2.5, p = 0.11). Females had insignificantly longer VOT duration for the Palestinian Arabic /t/ sound (M = 40, SD = 14) than males ( M = 33.3, SD = 13.1).

Furthermore, the effect of English phonetic instruction on VOT of the Palestinian Arabic voiceless plosive /t/ was not statistically significant (F (1, 38) = 0.16, p = 0.68). The group who received English phonetic instruction had insignificantly shorter VOT duration for the Palestinian Arabic /t/ (M = 35.5, SD = 14.7) than the group who did not receive any English phonetic instruction (M = 38, SD = 13.3) (see figure 9).

(25)

Figure 9 shows the mean VOT duration of the Arabic voiced /t/ for the four groups.

There was no significant interaction between gender and English phonetic instruction in the pronunciation of the Arabic /t/, (F (1, 38) = 1.07, p = 0.30) (see figure 10).

Figure 10 shows the interaction between groups in pronouncing the Palestinian Arabic voiceless plosive /t/.

5.1.3 VOT of Palestinian Arabic /k/

In this part of the first experiment, it was found that the Palestinian Arabic voiceless plosive /k/ has positive VOT duration (M = 48, SD = 15). This means that on average, the participants had the burst of the voiceless plosive /k/ in monosyllabic word initial position around 48 ms earlier to the vibration of their vocal folds when this plosive occurs before the long vowel /i:/ in nucleus position.

36.5 38 35.5 33.3 40 0 20 40 60 80 100 Average No phonetic instruction Phonetic instruction Males Females

Arabic /ti:/ VOT duration (ms)

(26)

45.4 41.1 52 53.3 0 10 20 30 40 50 60 70 80 90 100 Yes No A rab ic / ki :/ VOT (m s) Phonetic instruction Males Females

A two-way ANOVA test showed that the effect of gender on VOT of the Palestinian Arabic /k/ sound was statistically significant (F (1, 38) = 4.29, p < 0.05). On average, females had significantly longer VOT duration for the Palestinian Arabic voiceless plosive /k/ (M = 52.7, SD = 14.6) than males (M = 43.2, SD = 14.3).

However, the effect of English phonetic instruction on VOT of the Palestinian Arabic /k/ was not statistically significant (F (1, 38) = 0.08, p = 0.76). The group who received English phonetic instruction had very close results for the VOT of the Palestinian Arabic /k/ (M = 48.7, SD = 16) to the group who did not receive any English phonetic instruction ( M = 47.3, SD = 14.4) (see figure 11).

Figure 11 shows the mean VOT duration of the Arabic voiced /k/ for the four groups.

There was no significant interaction between gender and English phonetic instruction in the pronunciation of the Arabic /k/, (F (1, 38) = 0.41, p = 0.52) (see figure 12). 48 47.3 48.7 43.2 52.7 0 20 40 60 80 100 Average No phonetic instruction Phonetic instruction Males Females

Arabic /ki:/ VOT duration (ms)

(27)

In this experiment, the hypothesis that all groups will have lead voicing VOT for the Arabic voiced plosive /b/ was confirmed (3 & the first part of 1 in table 3). On the other hand, the second part of the first hypothesis that the voiceless plosives /t/ and /k/ belong to short lag VOT category was refuted. In addition, the hypothesis that females will have shorter lead voicing duration for the voiced plosive /b/ and longer VOT duration for the voiceless plosive /k/ plosives was confirmed (2), but the difference was not significant regarding the voiceless plosive /t/. Furthermore, the hypothesis that there is no significant difference between phonetic instruction groups regarding the VOT of Palestinian Arabic plosives /b/, /t/, and /k/ was found to be true (4).

Table 6. A summary of the statistical results for the first experiment (Palestinian Arabic pronunciation).

Palestinian Arabic pronunciation Df1 Df2 F p

/b/ Gender 1 38 7.96 0.008 Phonetic instruction 1 38 0.89 0.35 Gender*Phonetic instruction 1 38 0.62 0.43 /t/ Gender 1 38 2.50 0.11 Phonetic instruction 1 38 0.16 0.68 Gender*Phonetic instruction 1 38 1.07 0.30 /k/ Gender 1 38 4.29 0.04 Phonetic instruction 1 38 0.08 0.76 Gender*Phonetic instruction 1 38 0.41 0.52

5.2 VOT of English Plosives in Authentic Pronunciation

This experiment aims at investigating the pronunciation of the English plosives /p/, /t/, and /k/ by Palestinian EFL learners and the effect of gender and English phonetic instruction on VOT of these plosives. A two-way ANOVA test was carried out for each of these plosives to see if there is a difference between the gender groups and phonetic instruction groups, and to see if there is an interaction between them. 5.2.1 VOT of English /p/ in Authentic Pronunciation

A two-way ANOVA test showed that the effect of gender on VOT of the English voiceless plosive /p/ in participants' authentic pronunciation was not statistically significant (F (1, 38) = 0.047, p = 0.82). The VOT mean duration of females (M = 22.0, SD = 33.7) was very similar to VOT mean duration of males (M = 23.0, SD = 38.5).

(28)

average, the group who received English phonetic instruction had significantly longer VOT duration for the English voiceless plosive /p/ (M = 34.8, SD = 22) than the group who did not receive any English phonetic instruction (M = 10, SD = 42) (see figure 13).

Figure 13 shows the mean VOT duration of the English voiceless plosive /p/ for the four groups in authentic English pronunciation.

No significant interaction between gender and English phonetic instruction was found in the pronunciation of the English /p/ (F (1, 38) = 0.42, p = 0.51) (see figure 14).

Figure 14 shows the interaction between groups in pronouncing the English voiceless plosive /p/ in authentic English pronunciation.

ن 10 34.8 23 22 0 20 40 60 80 100 No phonetic instruction Phonetic instruction Males Females

Authentic English /pi:/ VOT duration (ms)

(29)

5.2.2 VOT of English /t/ in Authentic Pronunciation

A two-way ANOVA test showed that the effect of gender on VOT of the English voiceless plosive /t/ in the participants' authentic pronunciation was not statistically significant (F (1, 38) = 1.138, p = 0.29). Females had insignificantly longer VOT duration for the English voiceless plosive /t/ (M = 49.2, SD = 11.4) than males (M = 45.3, SD = 13.9).

Also, the two-way ANOVA test showed that the effect of phonetic instruction on VOT of the English voiceless plosive /t/ in the participants' authentic pronunciation was not statistically significant (F (1, 38) = 0.166, p = 0.68). The group who received English phonetic instruction had very similar VOT duration for the English voiceless plosive /t/ (M = 47.8, SD = 13.9) to the group who did not receive any English phonetic instruction (M= 46.8, SD = 11.8) (see figure 15).

Figure 15 shows the mean VOT duration of the English voiceless plosive /t/ for the four groups in authentic English pronunciation.

There was no significant interaction between gender and English phonetic instruction in the pronunciation of the English /t/ (F (1, 38) = 3.15, p = 0.085) (see figure 16). 46.8 47.8 45.3 49.2 0 20 40 60 80 100 No phonetic instruction Phonetic instruction Males Females

(30)

Figure 16 shows the interaction between groups in pronouncing the English voiceless plosive /t/ in authentic English pronunciation.

5.2.3 VOT of English /k/ in Authentic Pronunciation

A two-way ANOVA test showed that the effect of gender on VOT of the English voiceless plosive /k/ in the participants' authentic pronunciation was not statistically significant (F (1, 38) = 0.004, p = 0.95). Females had similar VOT duration for the English voiceless plosive /k/(M = 62.1, SD = 12.5) to males (M = 62, SD = 15.1).

(31)

Figure 17 shows the mean VOT duration of the English voiceless plosive /k/ for the four groups in authentic English pronunciation.

No significant interaction between gender and English phonetic instruction was found in the pronunciation of the English /k/, (F (1, 38) = 2.3, p = 0.13) (See figure 18).

Figure 18 shows the interaction between groups in pronouncing the English voiceless plosive /k/ in authentic English pronunciation.

In this experiment, the hypothesis that all groups will have short lag VOT for the English voiceless plosives in authentic English pronunciation were refuted (5 & 7 in table 3); short lag VOT was found only for the English voiceless plosive /p/, and the only group that did not produce short lag VOT for this plosive was the group who received English phonetic instruction. In addition, the hypothesis that females will have longer VOT duration for the English voiceless plosives than males was refuted

61.3 63 62 62.1 0 20 40 60 80 100 No phonetic instruction Phonetic instruction Males Females

Authentic English /ki:/ VOT duration (ms)

(32)

(6). Finally, the first part of the hypothesis that the group who received English phonetic instruction will have significantly longer VOT duration for the voiceless plosive /p/ was confirmed (8). However, the second part of the hypothesis that the group who received English phonetic instruction will have longer VOT duration for the English voiceless plosives /t/ and /k/ was refuted.

Table 7. A summary of the statistical results for the second experiment (authentic English pronunciation).

Authentic English pronunciation Df1 Df2 F p

/p/ Gender 1 38 0.047 0.82 Phonetic instruction 1 38 5.4 0.026 Gender*Phonetic instruction 1 38 0.42 0.51 /t/ Gender 1 38 1.13 0.29 Phonetic instruction 1 38 0.16 0.68 Gender*Phonetic instruction 1 38 3.15 0.085 /k/ Gender 1 38 0.004 0.95 Phonetic instruction 1 38 0.123 0.72 Gender*Phonetic instruction 1 38 2.3 0.13

5.3 VOT of English Plosives in Converged Pronunciation

This experiment aims at investigating the participants' pronunciation after being merely asked to repeat the English words they heard by a male English native speaker, and to see which gender and phonetic instruction groups converge more to English native speakers. A two-way ANOVA test was conducted for each English plosive to see if there is an overall difference between the groups after convergence. In addition, to compare between the groups in authentic English pronunciation and converged English pronunciation of the English voiceless plosives /p/, /t/, and /k/, the mean VOT duration of the English plosives in authentic pronunciation was subtracted from the mean VOT duration in converged pronunciation, and a two-way ANOVA test was carried out to see which group showed more convergence to a male English native speaker’s pronunciation.

(33)

Figure 19 shows the VOT duration of /p/, /t/, and /k/ of male English native speakers when pronouncing words in Table 4.

5.3.1 VOT of English /p/ in Converged Pronunciation

A two-way ANOVA test showed there was no significant difference in VOT of the English voiceless plosive /p/ between females and males after convergence (F (1, 38) = 0.324, p = 0.57). Females had insignificantly shorter VOT duration for the English voiceless plosive /p/ in converged English pronunciation (M = 37.7, SD = 26) than males (M = 44.2, SD = 28.5).

(34)

Figure 20 shows the mean VOT duration of the English voiceless plosive /p/ for the four groups in converged English pronunciation.

No significant interaction between gender and English phonetic instruction was found in the converged pronunciation of the English /p/, (F (1, 38) = 0.35, p = 0.55) (See figure 21).

Figure 21 shows the interaction between groups in pronouncing the English voiceless plosive /p/ in converged English pronunciation.

Another two-way ANOVA test which investigated the means after subtracting the mean VOT duration of the English voiceless plosive /p/ in authentic English pronunciation from the mean VOT duration of the same plosive in converged English pronunciation showed that the difference between females and males in the

34.6 47.3 44.2 37.7 0 20 40 60 80 100 No phonetic instruction Phonetic instruction Males Females

Converged English /pi:/ VOT (ms)

(35)

convergence to a male English native speaker's pronunciation of the English voiceless plosive /p/ was not statistically significant (F (1, 38) = 0.60, p = 0.44). Males showed more but insignificant convergence to a male English native speaker’s pronunciation of the English voiceless plosive /p/ (M = 20.5, SD = 28.5) than females (M = 15.8, SD = 26.8).

Furthermore, the two-way ANOVA test showed that the difference between the group who received English phonetic instruction and the group who did not receive any English phonetic instruction in the convergence to a male English native speaker's pronunciation of the English voiceless plosive /p/ was not statistically significant (F (1, 38) = 2.7, p = 0.10). The group who did not receive English phonetic instruction showed more but insignificant convergence to a male English native speaker’s pronunciation of the English voiceless plosive /p/ (M = 24.7, SD = 34.1) than the group who received English phonetic instruction (M = 11.8, SD = 17) (See figure 22).

Figure 22 shows the mean VOT duration difference between converged English pronunciation and authentic English pronunciation of the English voiceless plosive /p/.

No significant interaction between gender and English phonetic instruction was found, (F (1, 38)= 0.024, p = 0.85) (See figure 23).

24.7 15.8 20.5 15.8 0 10 20 30 40 50 60 70 80 90 100 No phonetic instruction Phonetic instruction Males Females

(36)

Figure 23 shows the interaction between groups after calculating the mean VOT duration difference between authentic and converged English pronunciation of the English voiceless plosive /p/.

5.3.2 VOT of English /t/ in Converged Pronunciation

A two-way ANOVA test showed there was no significant difference in VOT of the English voiceless plosive /t/ between females and males after convergence (F (1, 38) = 0.125, p = 0.72). Females had insignificantly shorter VOT duration for the English voiceless plosive /t/ in converged English pronunciation (M = 43.6, SD = 9.9) than males (M = 46.5, SD = 20.5).

A two-way ANOVA test showed there was no significant difference in VOT of the English voiceless plosive /t/ between the group who received English phonetic instruction and the group who did not receive any English phonetic instruction after convergence (F (1, 38) = 2.88, p = 0.09). The group who received English phonetic instruction had insignificantly longer VOT duration for the English voiceless plosive /t/ in converged English pronunciation (M = 49.3, SD = 23.8) than the group who did not receive English phonetic instruction (M = 40.8, SD = 13.6) (See figure 24).

(37)

Figure 24 shows the mean VOT duration of the English voiceless plosive /t/ for the four groups in converged English pronunciation.

No significant interaction between gender and English phonetic instruction was found (F (1, 38) = 1.6, p = 0.20) (See figure 25).

Figure 25 shows the interaction between groups in pronouncing the English voiceless plosive /t/ in converged English pronunciation.

To compare between the mean VOT duration of the groups in authentic English pronunciation and converged English pronunciation of the English voiceless plosive /t/, the mean VOT duration of the English sound /t/ in authentic pronunciation was subtracted from the mean VOT duration in converged pronunciation, and a two-way ANOVA test was carried out. The two-way ANOVA test showed that the difference between females and males in the convergence to a male English native speaker's pronunciation of the English voiceless plosive /t/ was not statistically significant (F (1,

Converged English /ti:/ VOT (ms)

(38)

38) = 1.93, p = 0.17). Males (M = 1.3, SD = 17.1) and females (M = -5.9, SD = 11.3) showed no significant difference between authentic and converged English pronunciation of the English voiceless plosive /t/.

Moreover, the two-way ANOVA test showed the difference between the group who received English phonetic instruction and the group who did not receive any English phonetic instruction in the convergence to a male English native speaker's pronunciation of the English voiceless plosive /t/ was not statistically significant (F (1, 38) = 2.44, p = 0.12). The group who received English phonetic instruction (M = -6.2, SD = 14.2) and the group who not received any English phonetic instruction (M = 1.6, SD = 14.6) showed no significant difference between authentic and converged English pronunciation of the English voiceless plosive /t/ (See figure 26).

Figure 26 shows the mean VOT duration difference between converged English pronunciation and authentic English pronunciation of the English voiceless plosive /t/.

No significant interaction between gender and English phonetic instruction was found, (F (1, 38) = 0.34, p = 0.85) (See figure 27).

-6.2 1.6 1.3 -5.9 -50 -30 -10 10 30 50 No phonetic instruction Phonetic instruction Males Females

(39)

Figure 27 shows the interaction between groups after calculating the mean VOT duration difference between authentic and converged English pronunciation of the English voiceless plosive /t/.

5.3.3 VOT of English /k/ in Converged Pronunciation

A two-way ANOVA test showed there was no significant difference in VOT of the English voiceless plosive /k/ between females and males after convergence (F (1, 38) = 2.49, p = 0.12). Females had insignificantly shorter VOT duration for the English voiceless plosive /k/ in converged English pronunciation (M = 72.5, SD = 14) than males (M = 81, SD = 20.3).

(40)

Figure 28 shows the mean VOT duration of the English voiceless plosive /k/ for the four groups in converged English pronunciation.

No significant interaction between gender and English phonetic instruction was found (F (1, 38) = 3.5, p = 0.07) (See figure 29).

Figure 29 shows the interaction between groups in pronouncing the English voiceless plosive /k/ in converged English pronunciation.

To compare between the mean VOT of the groups in authentic English pronunciation and converged English pronunciation of the English voiceless plosive /k/, the mean VOT duration of the English sound /k/ in authentic pronunciation was subtracted from the mean VOT duration in converged pronunciation, and a two-way ANOVA test was carried out. The two-way ANOVA test showed that the difference between females and males in the convergence to a male English native speaker's pronunciation of the English voiceless plosive /k/ was not statistically significant (F

75 79 81 72.5 0 20 40 60 80 100 No phonetic instruction Phonetic instruction Males Females

Converged English /ki:/ VOT (ms)

(41)

(1, 38) = 3.36, p = 0.074). Males showed more but insignificant convergence to a male English native speaker’s pronunciation of the English voiceless plosive /k/ (M = 19.4, SD = 15.8) than females (M = 10.4, SD = 14.5).

Furthermore, the two-way ANOVA test showed the difference between the group who received English phonetic instruction and the group who did not receive any English phonetic instruction in the convergence to a male English native speaker's pronunciation of the English voiceless plosive /k/ was not statistically significant (F (1, 38) = 0.073, p = 0.78). The group who received English phonetic instruction showed more but insignificant convergence to a male English native speaker‘s pronunciation of the English voiceless plosive /k/ (M = 16.2, SD = 14.6) than the group who did not receive English phonetic instruction (M = 13.6, SD = 16.9) (See figure 30).

Figure 30 shows the mean VOT duration difference between converged English pronunciation and authentic English pronunciation of the English voiceless plosive /k/.

No significant interaction between gender and English phonetic instruction was found (F (1, 38) = 0.55, p = 0.45) (See figure 31).

(42)

Figure 31 shows the interaction between groups after calculating the mean VOT duration difference between authentic and converged English pronunciation of the English voiceless plosive /k/.

In this experiment, the hypothesis that all groups will show convergence to a male English native speaker's pronunciation of the English voiceless plosive /p/ was confirmed (the first part of 9 & 11). However, the second part of hypotheses that gender groups and phonetic instruction groups will not show convergence to a male English native speaker's pronunciation of the English voiceless plosives /t/ and /k/ was refuted; all groups showed convergence to the pronunciation of /k/ but not /t/. Furthermore, the hypothesis that females will still have longer VOT duration for the English voiceless plosives than males after convergence was refuted (10); although males had longer VOT duration for the English voiceless plosives, the difference was not significant. In addition, the hypothesis that the group who received English phonetic instruction will have longer VOT duration for the English voiceless plosives after convergence was refuted (12); although the group who received English phonetic instruction had longer VOT duration, the difference was not significant.

(43)

Table 8. A summary of the statistical results for the third experiment (converged English pronunciation).

Converged English pronunciation Df1 Df2 F p

/p/ Gender 1 38 0.324 0.57 Phonetic instruction 1 38 1.96 0.16 Gender*Phonetic instruction 1 38 0.35 0.55 /t/ Gender 1 38 0.125 0.72 Phonetic instruction 1 38 2.88 0.09 Gender*Phonetic instruction 1 38 1.6 0.20 /k/ Gender 1 38 2.49 0.12 Phonetic instruction 1 38 0.24 0.62 Gender*Phonetic instruction 1 38 3.5 0.07

Table 9 A summary of the statistical results after subtracting the mean VOT duration of the English voiceless plosives /p/, /t/, and /k/ in authentic English pronunciation from the mean VOT duration of these plosives in converged English pronunciation.

Converged-Authentic English pronunciation Df1 Df2 F p

(44)

6. Discussion

A review of the results is illustrated in the figure bellow:

Figure 32 shows a visual illustration of the results from the three experiments (the figure does not present the exact position of the sounds as it is used to show the differences between groups).

6.1 Palestinian Arabic

6.1.1 The VOT Category of Palestinian Arabic Plosives

In the literature review section, I have introduced two studies regarding the VOT of Arabic plosives. In the first study, Aldahri (2012) found that Arabic voiced plosives, including the /b/ sound, belong to the short lag VOT category while Arabic voiceless plosives belong to the long lag VOT category. On the other hand, Khattab (2012) found that Arabic voiced plosives belong to the lead voicing VOT category, whereas Arabic voiceless plosives belong to the short lag VOT category.

(45)

that have a VOT duration between 30 ms and 50 ms. Therefore, I suggest the paradigm presented in figure 33 for the Palestinian Arabic plosives /b/, /t/, and /k/.

Figure 33 shows the VOT categories to which the Palestinian Arabic plosives /b/, /t/, and /k/ belong.

Still, I assume that my findings are more similar to Khattab's (2012) for two possible reasons. The first reason is attributed to the validity of the methodology in Aldahri's (2012) study in which he compares between the VOT of MSA (Modern Standard Arabic) and CA's (Classical Arabic) plosives. The only difference between MSA and CA is style and vocabulary, but none of them is spoken on a daily basis as a local dialect anywhere. CA is a literary language which dates back to the 6th century, and it is the chosen language of the Quran, whereas MSA is the standard Arabic language that is used in formal occasions like education, journalism, and modern literature (Ryding 2011). In his study, Aldahri does not mention anything about the participants' background, where they are from, how they know the CA, or if they are versed in the MSA. Even if the participants were proficient in CA and MSA, there would be no guarantee that they would not transfer the plosives from their local dialect.

(46)

6.1.2 Males vs. Females

In this study, it was found that there is a significant effect of gender on the VOT duration of the Arabic plosives /b/ and /k/. However, no significant effect of gender on the VOT duration of the voiceless plosive /t/ was found (see figures 7, 9, & 11). It was found that males had significantly longer lead voicing duration before the release of the closure of the voiced plosive /b/ than females. According to van Alphen and Smits (2004), the reason why females have shorter lead voicing VOT for the voiced plosives than males is that females have shorter vocal tract size "Assuming equal volume velocity through the glottis, oral pressure will tend to rise more quickly in females than in males, which makes it harder to produce prevoicing" (p.459).

(47)

"There are two ways of considering this. Firstly, the cavity behind the velar stop has a smaller volume than that behind the alveolar or bilabial stops. Secondly, the cavity in front of the velar stop has a larger volume than that in front of the alveolar or bilabial stops"(Cho and Ladefoged 1999: 209).

Since females have shorter vocal tract than males, which also means smaller cavity behind the velar plosive, the effect of gender on the VOT duration might be significant for the voiceless plosive /k/ but not /t/.

6.1.3 English Phonetic Instruction vs. No English Phonetic Instruction

It was found that the effect of English phonetic instruction on VOT of Palestinian Arabic plosives was not significant; both groups had similar VOT duration for the Arabic plosives /b/, /t/, and /k/ which suggests that there is no interference from L2 to L1. In other words, studying the pronunciation of English plosives does not have an effect on the pronunciation of Arabic plosives (see figures 7, 9, & 11).

6.2 Authentic English Pronunciation

6.2.1 Females vs. Males

In this study, it was found that both gender groups produced a short lag VOT for the English voiceless plosive /p/ as predicted in the hypotheses section due to L1 interference. However, no significant difference between males and females was found; both groups produced around 22ms VOT duration for this sound (see figure 13). In addition, the participants produced an "in between" VOT for the voiceless plosive /t/ just like Arabic but with a little longer VOT duration. This suggests that the participants do not assume that there is a difference between the Palestinian Arabic /t/ and the English /t/ (see figure 15). However, the participants in both groups produced a long lag VOT in the pronunciation of the voiceless plosive /k/. The repetition of the same plosive for a second time after the first experiment might have caused the participants to have longer VOT duration for this plosive, which may also explain why the participants had longer VOT for the voiceless plosive /t/. The average VOT duration of the Arabic /k/ is 48 ms, which I have assumed to be "in between" category, is very close to the long lag category, and with a little increase in the VOT duration it crossed the long lag VOT boundary with a mean duration of around 62ms.

The Pronunciation of English Voiceless Plosives /p/, /t/, and /k/ by Palestinian EFL Learners.

Master Thesis