ANALYSIS OF PRONUNCIATION ERRORS OF SAUDI ESL LEARNERS by Turki A. Binturki B.A., King Saud University 2001 A Thesis Submitted in Partial Fulfillment of the Requirements for the Master of Arts Degree Department of Linguistics in the Graduate School Southern Illinois University Carbondale August 2008 1456863 1456863 2008 THESIS APPROVAL ANALYSIS OF PRONUNCIATION ERRORS OF SAUDI ESL LEARNERS By Turki A. Binturki A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Arts in the field of Applied Linguistics and Teaching English to Speakers of Other Languages Approved by: Karen Baertsch, PhD, Chair Krassimira Charkova, PhD Laura Halliday, PhD Graduate School Southern Illinois University Carbondale August 2, 2008 i AN ABSTRACT OF THE THESIS OF Turki A. Binturki, for the Master of Arts degree in Applied Linguistics and Teaching English to Speakers of Other Languages, presented on May 30, 2008, at Southern Illinois University Carbondale. TITLE: ANALYSIS OF PRONUNCIATION ERRORS OF SAUDI ESL LEARNERS MAJOR PROFESSOR: Dr. Karen Baertsch This study was conducted as a qualitative investigation to determine the difficulty of /p v / to Saudi ESL speakers. More specifically, this study investigated what word environments were most difficult for Saudi speakers. Subjects were specifically selected to represent Najdi dialect because it is the closest dialect to Classical Arabic and also for the lack of previous research regarding this dialect. Five native speakers of Saudi Arabian Najdi dialect studying in the U.S. were chosen to participate in this study. A word list and a reading passage were used to elicit the target sounds in order to generate data for both context and isolation. The results of this study demonstrate that Saudi ESL speakers do have difficulty with the voiced interdental fricative /v/ and to lesser extent, with /p/ and //. The study also found that difficulty ii was closely related to certain word positions. The findings of this study are important to teachers of English as a second language especially those who deal with Saudi speakers. iii DEDICATION I would like to dedicate this thesis to my dear mom, dad, and my brothers who have always supported me. iv ACKNOWLEDGMENTS This thesis would not be possible if supervision did not exist. For that, I am extremely grateful to Dr. Baertsch for being a mentor and a phenomenal linguist. Her energy and dedication to linguistics inspired me to seek new levels of education. Her knowledge of linguistics is highly valued and for her support during my masters program was unprecedented from teachers. I would also like to thank my committee members Dr. Charkova and Dr. Halliday. They both were more than teachers to me. They taught me the meaning of research and seeking information and most importantly pulling all that off with a smile. Special thanks Brett A. Burkardt who is a true friend. I could not imagine going through this degree program without him. Also huge thanks to Amanda Coyne for transcribing sounds and to her endless comfort. Many thanks to Diane Korando for her support. Also, thanks to the subjects who took part in this study. Last but certainly not least huge thanks to my wife Ghada for her endless patience and support. She is my perfect balance. v TABLE OF CONTENTS CHAPTER PAGE ABSTRACT ........................................................................................................................ i DEDICATION ...................................................................................................................iii ACKNOWLEDGMENTS ................................................................................................. iv LIST OF TABLES ............................................................................................................ vii LIST OF FIGURES ........................................................................................................ viii CHAPTERS CHAPTER 1 – Introduction .................................................................................. 1 CHAPTER 2 – Literature Review ...................................................................... 24 CHAPTER 3 – Methodology .............................................................................. 35 CHAPTER 4 – Results......................................................................................... 45 CHAPTER 5 – Discussion, Implications, Limitations and Conclusion........... 60 REFERENCES .................................................................................................................. 73 APPENDICIES Appendix A ........................................................................................................... 81 Appendix B .......................................................................................................... 82 vi Appendix C .......................................................................................................... 83 Appendix D .......................................................................................................... 84 VITA ................................................................................................................................ 85 vii LIST OF TABLES TABLE PAGE Table 2.1 ............................................................................................................................ 20 Table 4.1 ............................................................................................................................ 46 Table 4.2 ............................................................................................................................ 48 Table 4.3 ............................................................................................................................ 49 Table 4.4 ............................................................................................................................ 50 Table 4.5 ............................................................................................................................ 52 Table 4.6 ............................................................................................................................ 53 Table 4.7 ............................................................................................................................ 54 Table 4.8 ............................................................................................................................ 55 Table 4.9 ............................................................................................................................ 57 Table 4.10 .......................................................................................................................... 58 Table 4.11 .......................................................................................................................... 59 viii LIST OF FIGURES FIGURE PAGE Figure 4.1 ........................................................................................................................... 53 Figure 4.2 ........................................................................................................................... 56 Figure 4.3 ........................................................................................................................... 58 1 CHAPTER 1 INTRODUCTION Many second language learners want to “sound like natives”, however this dream has yet to be achieved by the majority of second language learners are under the impression that they speak English intelligibly, but the truth is a vast majority of ESL learners make numerous errors. In addition to phonological, lexical and spelling errors, learners also produce pronunciation, rhythm, intonation voice quality errors (Messiha, 1985). To a greater extent, ESL learners are not even aware of the pronunciation errors that they make. There are numerous factors that contribute to the errors second language learners produce, especially in the area of pronunciation. Certainly, the differences between the learners‟ native language and the acquired language play an important role in these difficulties. Moreover, there are elements that contribute directly to pronunciation errors. In addition to variations in the phonemic inventory of both languages, age and language transfer also play a strong role in the acquisition process. 2 1.1 Attitudes towards pronunciation English occupies a high status among world international languages as it has become the language of diplomacy, trade, communication, technology, business and development of large numbers of Arab countries (Suleiman, 1993). Its importance has shaped the positive attitudes of some second language learners. Suleiman (1993) stated that according to the participants in his study; learning English provides the individual an advantage as an active participant in today‟s world, opening new horizons to a better future. A large population of second language learners believes that the main difficulty they experience when speaking English is pronunciation. Generally, they consider pronunciation as the main cause for their communication problems (Derwing & Rossiter, 2002; Al-Kahtany, 1995). The majority of ESL learners want to be indistinguishable from English native speakers. In a study conducted by Derwing (2003), more than half of the interviewees felt that they would receive more respect if they pronounced English well. According to (Giles, 1970; Weener, 1967), it is very important to second language learners to be perfectly understood and not be stigmatized as being “lower class” and have anything 3 they produce be discounted by others. Good pronunciation insures that a message gets transmitted more effectively. Also, abnormal pronunciation means that the person is stained as a “non-native” speaker of that language (Weeren & Theunissen, 1987). Often, native speakers of English evaluate non-native speakers of English negatively. Ryan and her colleagues found out that U.S. listeners of English negatively rated Spanish and German accented English speech (Ryan, Carranza & Moffie, 1977; Ryan & Sebastian, 1980; Ryan & Bulik, 1982). However, not all non-native speakers of English are stigmatized due to their accents. In a study conducted by Cargile & Giles (1998), a Japanese speaker with a “moderate accent” was rated highly by American university students. There is a large number of learners that attribute communication problems to pronunciation, as a result this conception reflects negatively on their attitude towards their target languages. Listeners and speakers are aware of their communication problem, thus they are interested in improving their speaking skills. This study identifies which pronunciation problems to expose so that educators and English teachers can more effectively deal with them. 4 In the following section the predominant issues that hinder the acquisition process of many Saudi Speakers learning English as a second language will be discussed. 1.2 Difficulties of English Acquisition. Non-native speakers of English encounter endless difficulties on their mission to acquire a second language. Some of these difficulties are phonological, syntactic, lexical, social, attitudinal and communicative. A New York corporation discovered that communication difficulties between native and non-native speakers of English was limiting the company‟s efficiency (Mettler, 1989). Differences between languages can cause difficulties in acquiring a second language. Many differences appear when cross-comparing two languages. For example, when comparing English to Cantonese, we see differences in the phonemic inventory, characteristics of sounds, distribution of phonemes, syllable structure and rhythm (Chan & Li, 2000). On a phonological and structural level, English is very different from Mandarin Chinese. Chan & Li (2000) expected that ESL Chinese learners would meet several difficulties in their acquisition process. 5 In our first language experience, it seems that we are able to process native language with ease despite the variability in speaker sounds, accents, speech rate and emotional affect (Zhang & Wang, 2007). Pallier, Christophe, & Mehler (1997) explained that listeners of different languages use their perceptual system to exploit knowledge about the constraints on the co-occurrence of phonemes. Pallier & his colleagues (1997) added that speakers predict the upcoming phoneme using phonotactics. They also attempt to establish a pattern of the speech signal that follows the patterns of their native language. Japanese speakers insert “illusory” phonemes when a word does not conform to this pattern. As a result, speakers of Japanese (a language that does not allow word-internal obstruent clusters) have a lot of trouble discriminating between VCCV and VCVCV. Another difficulty learners of English encounter is L1 interference. Mukattash (1981) attributed errors of the auxiliary verb BE replacing other auxiliaries in 400 first year students at the University of Jordan to L1 interference. He added that L1 interference played a more significant role on the phonological and the lexical levels than the syntactic and morphological levels. Moreover, he explained that the role was clear in “style”, especially in written complex sentences as opposed to simple sentences. 6 One of the goals of this study is to introduce pronunciation errors committed by Saudi ESL learners in order to identify the source or sources of difficulty. 1.3 The Markedness Differential Hypothesis (MDH). It is well known that Eckman (1977) proposed the Markedness Differential Hypothesis (MDH) as justification for areas of difficulties in second language learning. He proposed his hypothesis on the phonological theory of markedness. The common sounds in many languages are considered unmarked, while the less common sounds are considered marked. Eckman predicted that for second language learners, the acquisition of an unmarked sound like unaspirated stops would be easier than the aspirated stops which are considered marked (Fellbaum, 1996). The MDH has gathered a wide range of audiences, some of whom agree with it and think it is the cause of second language errors, while others think it cannot be the sole answer to identify sources of errors. The current study identifies pronunciation errors which could lead to predicting sources of difficulty. 7 1.4 The Critical Period Hypothesis (CPH). Age has received its fair amount of attention and research as a controversial major factor in language acquisition. More specifically, the Critical Period Hypothesis (CPH) is still disputed in many age related language learning studies. Although age is irrelevant to the current study, it is cited enough in the literature so that it must also be discussed here. Chiswick & Miller (2007) defined the CPH as a sharp decline in learning outcome with age. They added that to ensure a native-like proficiency, one has to acquire language before the critical period ends (at puberty). Learning a second language is also much more difficult after the critical period. Lenneberg (1967) established that language learning ability would be affected after the age of puberty due to the completion of lateralization processes. Krashen (1973) also suggested that the period affected more than general learning abilities covering other areas like pronunciation, morphology, syntax, or pragmatics in language. In response to Lenneberg‟s hypothesis, (language learning was possible during the period from infancy to puberty, with a loss of abilities after puberty), Johnson & Newport (1989) concluded that the data on second language learning in their study, did 8 carry a corresponding result. They did not find a direct relationship between performance and age of learning throughout childhood, with a rapid decline in performance marking the end of the critical period; rather, in their study performance increasingly declined from about age seven until adulthood. As mentioned earlier, age has no significance to this study because all of the subjects began their acquisition of English after the critical period. Nevertheless, the role of age in learning languages has long been debated. It was worthy of referring to the effects of age on second language learning. 1.5 Types of Errors There are errors committed by language learners in every area of linguistics. Tushyeh (1996) exemplified multiple errors committed by Arab learners of English on various linguistic areas. On the phonological level, Arab learners of English confuse /p/ and /b/; /f/ and /v/; /i/ and /e/. On the morphological level, the same learners commit typical errors such as: *womens and *furnitures. She noted a lexical error example of: *“the clock is now ten a.m.” For the syntactic level she gave an example of wrong word order and the use of presumptive pronouns. Because the focus of this thesis is on 9 pronunciation difficulties of Arab learners of English, my focus in this section will be limited to pronunciation rather than other aspects of acquisition. A brief description of the types of errors is provided under this subtitle. Distinguishing between types of errors is significant because not all errors are dominant. Some errors proved to be temporal and developmental and some are fossilized and dominant. Corder (1967: 25) in his widely cited article, The Significance of Learner‟s Errors, first introduced a hypothesis that stated “learner‟s errors provide evidence of the system of the language that he is using at a particular point in the learning course”. He further classified errors into two types. The first type is mistakes. Mistakes are infrequent and non-systematic occurrences that may occur in the native language of a speaker. Errors are the second type. They are systematic occurrences in the use of the L2 as compared to the correct use of the native language (Bayraktaroglu, 1985). Burt & Kiparsky (1972) made a distinction between errors by classifying them into global and local. Global errors, unlike local errors, hinder communication. They provided some examples of both types; local errors as in the omission of third person singular, while an example of global errors would be the wrong order in a sentence. 10 Geranpayeh (2000) argued that learner‟s errors cannot be accounted for by a single theory. He proposed that a combination of contrastive analysis and analysis of learner‟s errors can be fruitful. He pointed out the differences between the English and Persian systems as the English system uses definite markers, while the Persian system uses specific markers. He also added that in English syntax plays a major role in the uses of definite markers, whereas in Persian, the semantic system is the major player in the specific markers. In his conclusion, he claimed that the acquisition of the English article system is delayed for the majority of L2 learners of Persian to the final stages of language learning (Geranpayeh, 2000). Predicting the systematic errors that ESL learners will produce is not impossible, but is a mere hypothesis. This was demonstrated by Bayraktaroglu (1985) in his error analysis study of native speakers of Turkish. He explained that the predictions of pronunciation errors made by Turkish subjects based on a contrastive analysis were not accurate. He added that the errors that were found were not predictable through a prior comparison of the English and Turkish systems, but by the analysis of the results of actual errors. 11 Breaking down errors helps better deal with a large problem such as the one at hand. Pronunciation errors are a prominent problem for learners and teachers of English alike. By identifying pronunciation errors and classifying them into subcategories, this will help to deal with the problem more efficiently. This study helps to accurately identify errors and provides better solutions to pronunciation problems. 1.6 Error Analysis and Contrastive Analysis There is growing interest in research regarding error analysis in the field of applied linguistics as opposed to contrastive analysis (Tushyeh, 1996). Error analysis (EA) is working with actual errors second language learners express. Contrastive analysis (CA) is comparing two languages to predict errors (Tushyeh, 1996). The belief that CA predicts areas of difficulties faced by learners of a foreign language has been questioned (Bayraktaroglu, 1985). The predictive powers of CA are often not completely accurate even in phonology. Corder (1967) suggested that the language system of an interlanguage is provided by the learner‟s errors. Corder indicated that this system is the one that the 12 learner has learned at a specific point despite the fact that the language system being produced is not yet correct. Bayraktaroglu (1985) compared the CA predictions of English and Turkish and the results he came up with in his EA study. In his study, clarified that CA predicted that Turkish learners of English may have difficulties in the labiodental fricatives [] and [] because of their absence from the Turkish phoneme inventory, but failed to accurately predict the produced sounds as well as following EA. A different representation for the problematic sounds were produced by Turkish informants where CA, could not have predicted (Bayraktaroglu, 1985). Bayraktaroglu (1985) concluded, after his comparison study of the two analysis systems, that error analysis cannot totally replace contrastive analysis. He added that it is safe to claim that EA might be supplemented by the results of CA especially when determining the sources of errors is required. The two methods of identifying errors are equally important. Contrastive analysis predicts where errors might lay in order for error analysis to accurately demonstrate them. This study is an error analysis study; it provides detailed errors committed by Saudi ESL learners. It also yields substitutions of the targeted sounds. 13 1.7 Sources of Difficulties There are two major sources of errors; interlanguage and intralanguage (Tushyeh, 1996). Many researchers, like Lado (1957) and Selinker (1972) focused on native language interference (interlanguage errors). According to Richards (1971), interlanguage is the cause of more than 36% of errors. Others like Dulay & Burt (1974) focused on the second language as the source of difficulty (intralanguage errors). Tushyeh (1996) listed numerous intralanguage errors; simplification, overgeneralization, hypercorrection, faulty teaching, avoidance, inadequate learning and false concepts hypothesized. Selinker (1972 p. 36) introduced the concept of fossilization as a persistent type of error that second language learners make. He defined it as “linguistic items, rules and subsystems which speakers of a particular native language will tend to keep in their interlanguage relative to a particular target language, no matter what the age of the learner or amount of explanation and instruction he receives in the target language”. He clarified his theory with the example of American English learners of French using the retroflex // instead of the French uvular //. 14 Another intralanguage error committed by second language learners is the overgeneralization of the target language rules, such as the overgeneralization of the past tense suffix (-ed) to irregular verbs which result in errors as in goed and comed (Tushyeh, 1996). Bayraktaroglu (1985) demonstrated that some errors made by Turkish learners of English were due to overgeneralization of the target language phonetic system. He attributed the substitutions of // by [] and // by [] as examples of overgeneralization of sounds which are not in the Turkish phonemic inventory. Furthermore, Turkish learners replace phonemes that exist in Turkish like /d/, /s/, /z/ and /t/ with newly acquired phones from English [] and []. Differences in writing systems are a continuous source of difficulty for Arab learners. Tushyeh (1996) explained that the Arabic alphabet is different from the Roman alphabet, where English has only twenty six characters, Arabic has twenty eight. Moreover, Arabic has more of a one-to-one correspondence sound system where Arabic letters correspond closely to phonemes. This language feature is different from English that has about 85% of its words in regular spelling, in addition to the exceptions of the pronunciation rules. On the other hand, Arabic writing system does not distinguish between orthography and pronunciation (Tushyeh, 1996). 15 Although the current study does not deeply investigate sources of problems that cause pronunciation errors, it attempts to figure out the reasons behind errors. 1.8 Pronunciation difficulties Yost (1959: 115) noted that Syrian and Lebanese learners of English often have difficulty distinguishing between certain sounds like [s] and [t]. For example, the word chestnuts is difficult for Arabic speakers. Syrian speakers of English would pronounce it like [gostnust] since [s] never follows [t] in the same syllable in Arabic. He also stressed the influence of Arabic sounds on learners‟ English by giving examples of vowel errors such as put pronounced as [put] and full turned into [ful]. In addition, Yost further explained the absence of some English phonemes in Arabic which were substituted by phonemes from the native language as in the following: /p/ became [b], /g/ was [k], /t/ had no Arabic equivalent but it wasn‟t difficult for educated people to pronounce, /v/ also caused little trouble and that probably was due to the vocalization of [f] as in the Arabic word mustashfa, /d/ became [] in words like good, // was pronounced [k], and finally /r/ was realized as a tongue pointed trill. 16 Learners of English face countless difficulties that hinder their learning process. Arab learners are no exception. Tushyeh (1996) stated that errors were crucial to the pronunciation of Arab learners of English. She reported errors caused by differences between the sound systems such as the substitution of the sound [b] for the letter p as in [bibəl] for people. She claimed that the Arabic sound system does not include a voiceless bilabial plosive, but only a voiced bilabial plosive /b/. It is important to highlight the difficulties in pronunciation that are specific to Arabic speakers and to not assume they will make the same errors as other English learners from other L1‟s which this study will attempt to accomplish. 1.9 Arabic dialects An interest in the Arabic language and Arabic phonetics was established since the 7th century CE. Al-Khalil and Sibawayh were the authors of the two first and far most important books in Arabic grammar, phonology and phonetics (Rosenhouse, 2007). This section will discuss differences among Arabic dialects. More specifically, it will demonstrate the differences between Saudi Arabian dialects. One of the reasons to establish a distinction between dialects is the fact that all Arabic dialects are spoken and 17 none are practiced in the written form (Al-Kahtany, 1997). This lack of a written language format may have some effect on pronunciation abilities and error production. Arabic in the Arab world, like many other languages, suffers from an existing situation of the separation of the al-rabia al-amia “colloquial Arabic” (regional dialects) or low variety and al-rabiyya al-fusa “literary Arabic” (Modern Standard Arabic) or high variety (Scott & Tucker, 1974; Al-Kahtany, 1997). This separation is known as diglossia. Gordon (2005) reported that Ethnologue includes 35 varieties of Arabic on its website. These versions of Arabic are scattered among 21 countries. These varieties of Arabic are not restricted to Arab countries, but they transcend borders to other nations including Afghanistan, Cyprus, Malta and Uzbekistan. In Saudi Arabia alone there are four dialects: Gulf, Hijazi, Najdi, and Standard spoken Saudi Arabian dialect. The highest and most prestigious form of Arabic language lies between the covers of the Quran (Mneimneh, 1997). The significance of Arabic lies in its early documentation which was in the 600s following the Islamic advent. Isaksson (1999) listed at least 45 different representations in his survey of non- standard first person singular pronouns in modern Arabic dialects. The diverse 18 representations were gathered by different scholars during different times with disparate linguistic backgrounds. It included data gathered from many countries and various tribes around the Arab world. Despite the many representations of one pronoun, Isaksson (1999) claimed that the separate pronoun (ana) has turned out to be very stable in modern Arabic dialects. Although Arabs in general share some pronunciation errors, there are errors that characterize some Arab nationalities more than others. Messiha (1985) clarified that among the pronunciation problems that Egyptian Arabic learners face are /p/, //, //, /v/ and these sounds were represented as follows [b], [s], [s], [f] respectively. This also corresponds with Barros‟ (2003) results with her Egyptian subjects. Messiha (1985) also pointed out that some Arab speakers do not share the same difficulties with Egyptian speakers regarding the interdental fricatives. This paper, will investigate pronunciation errors regarding specific ESL learners. More specifically, it will carefully look into three problematic consonant for Saudi ESL learners in the onset (beginning) and coda (end) environments of words and will focus on the phonemes / p v/. 19 The importance of this study lays in its involvement in exposing one of the Saudi dialects, a dialect that did not receive much attention in comparison to other dialects of the area. 1.10 American vs. Arabic sound systems Gordon (2005) lists the most extensive catalog of the world‟s languages. It contains a detailed classified list which currently includes 6,809 distinct languages (Anderson, 2005). English and Arabic languages are not close relatives. Where English is an Indo-European West Germanic language, Arabic is an Afro-Asiatic Central South Semitic language. Nevertheless, there are a lot of phonemes that are present in both systems (see table 2.1). English has twenty-five consonants /p b m n f v   t d s z h l   t d ɹ j k g   /, ten vowels /    u     ə/ and five diphthongs /aj aw j ej ow/ (Rogers, 2000). On the other hand, Arabic like many Semitic languages, is marked by a rich consonantal system and by a limited vocalic system. It has twenty-eight consonants /b d  f h j k l m n r s ʃ t  w z d  χ sˤ dˤ tˤ zˤ ʕ ɣ q ʔ/, six vowels / u  : u: i:/, and two diphthongs /i u/ (Kharma & Hajjaj, 1997, phonetic symbols 20 converted to standard IPA symbols where appropriate; see also Rogers 2000; Prochazka 1988; Al-Ani 1970; Watson 2002 ). Table 2.1 English vs. Arabic IPA chart (this table was reduced to fit the studied sounds) Bi la bi al La bi od en ta ls D en ta l A lv eo la r Po st al ve ol ar V el ar U vu la r Ph ar yn ge al G lo tta l Plosive Eng. p b t d k g  Arb. b t tˤ d dˤ k q  Fricative Eng. f v   s z   h Arb. f   s sˤ z zˤ  χ ɣ   h Trill Eng. Arb. r Approximant Eng.  Arb. Although there are many sounds that are shared by both sound systems, there are sounds that are present in one system but not the other. In this study, the focus is not the 21 vowel system, but the main objective is to concentrate on the consonantal sounds that are absent from Arabic language (assuming it to be the cause of some pronunciation difficulty). More specifically, Arabic lacks the voiceless bilabial stop /p/, voiced velar stop /g/, voiced fricative /v/, voiced postalveolar fricative //, voiceless post-alveolar affricate /t/, voiced nasal velar // and the voiced retroflex approximant //. Because covering all of these sounds is beyond the scope of this work, this list was further reduced to /p v / for the purposes of this study. These three particular sounds were chosen since they are well documented sounds in proving difficult for Arabic speakers of all dialects. While there are many sounds that cause pronunciation difficulties beyond these three sounds, this study focused only on these three sounds in order to have a smaller and more controlled study. It is very important to establish that the comparison mentioned above is between American English (AE) and Classical Arabic (CL) or Modern Standard Arabic (MSA). Furthermore, the dialect that all participants presented in this study was restricted to the Saudi Arabian Najdi dialect, which displays the majority of the characteristics that CL and MSA contain, including all the sounds CL and MSA Arabic lacks (Prochazka, 1988; Al-Ani, 1970; Watson, 2002; Kharma & Hajjaj, 1997). In addition, some of the 22 sounds that are absent from CL and MSA are present in some Arabic dialects such as // in Sarat abidah (Southern Hijaz and Tuhama dialect in Saudi Arabia) and Saida (Morocco), /v/ in Rufaidah (Southern Hijaz and Tuhama dialect in Saudi Araia) and Syrian Arabic, and /g/ in gilit (Mesopotamian dialect in Iraq) (Prochazka, 1988; Owens, 2005; Yoda, 2006 and Yost, 1959). Barros (2003) pointed out that Egyptian subjects reflected errors in pronunciation that were different from errors exhibited by other subjects from Saudi Arabia and Kuwait. These differences between dialects make it important to control subjects‟ dialect in any study of Arabic learners of English as shown in this study. 1.11 Purpose & Significance of the study One of the important skills that nearly all language learners want to master is pronunciation. Large numbers of L2 learners are concerned with their speaking abilities. Various studies have verified that ESL learners are not satisfied with their pronunciation skills. They always dream of speaking English fluently (Derwing & Rossiter, 2002). Each L1 speaker of a language produces some errors that are different from another L1 speaker of another native language. Scott & Tucker (1974) proposed that the 23 interference in spoken English by Arab learners is from colloquial Arabic. As the literature clarifies, certain speakers from Arabic dialects produce different errors than speakers of other dialects. This research highlights the significance of effects of dialects on pronunciation. The present study will look into the difficulties of pronunciation of Saudi Arabian ESL learners. More specifically, focusing on Najdi dialect including certain phonemes; the bilabial voiceless stop /p/, the labiodental voiced fricative /v/ and the alveolar approximant //. This study investigates which word positions Saudi ESL learners have the most difficulty with among the three targeted sounds. 24 CHAPTER 2 LITERATURE REVIEW The number of studies investigating L2 pronunciation is growing enormously. Each study is concerned with different aspects of pronunciation. Some studies deal with attitude towards pronunciation; others look into sources of difficulties in pronunciation. The following chapter will focus on pronunciation studies that have some significance to the present study. The literature will be presented in two main sections. The first will be studies that deal specifically with errors made by Arabic speakers. The second section of studies will demonstrate general difficulties made by learners of English whose first language is not necessarily Arabic and are included to demonstrate some common difficulties that are related to this study. Altaha (1995) focused on competence errors in pronunciation made by Saudi sophomores majoring in English. He collected data over a period of four years by recording sophomores in various situations. He identified seven major pronunciation errors. The first error reported was replacing the sounds /p, v, e/ and /ɛə/ with the sounds [b, f, i] and [əː], respectively. The second error that he recognized was vowel 25 insertion in words that contains a sequence of two consonants initially or inserting a vowel between the first two consonants in words that have a sequence of three consonants initially. The third identified error was the pronunciation of “silent” letters. Fourthly, pronouncing the letter c as [k] when it should be pronounced as [s]. The fifth error was pronouncing the voiced stop /g/ as a voiced affricate [dʒ] and vice versa. The sixth error was pronouncing the digraph dg as two phonemes medially, as in example 1: Ex. 1 [bʌd ɡit] for /bʌdʒət/ budget He finally identified shifting the place of stress as the seventh pronunciation error as in example 2: Ex. 2 [pri pɔziʃən] for /prepə ziʃən/ preposition This study is important because it demonstrates many of the English sounds that pose difficulty for Arabic speakers of English. Barros (2003) looked into pronunciation difficulties with English consonants facing ESL Arab learners after the age of puberty. She identified and analyzed difficulties in pronunciation of six Arabic speakers who have been living in the United States for four years. She explained that the sounds /ŋ/, /ɫ/, /d/, /p/, /ɻ/, /v/, /ʤ/ and /ð/ were mispronounced respectively as [k], [l], [t], [b], [r], [f], [ʒ] and [də]. 26 Barros reported that 5 out of 6 of the participants had problems with the following consonants (in order of difficulty): /ŋ/, /ɫ/, /d/, /p/ and /ɻ/. As for the sound /v/ only half of the participant had problems pronouncing it appropriately. The consonants /ʤ/ and /ð/ were mispronounced by 2 participants. Barros‟s subjects did not represent a single Arabic accent. As she pointed out, her subject selection was formed from different Arab countries. She explained that two subjects were from Egypt, one was from Saudi Arabia and three were from Kuwait. The two Egyptian subjects were the only participants who experienced problems with the phonemes /ʤ/ and /ð/. The other participants all displayed the same mispronunciations, but one subject from Kuwait who pronounced the sound /v/ appropriately. We can easily extract from this that there that there are differences in pronunciation among Arab learners. This gives us an insight that we cannot generalize pronunciation mistakes to all Arabs. There are certainly differences in various Arab backgrounds and dialects which result in different mispronunciations in English. Barros stated that the greatest number of participants in her study experienced problems with the phoneme /p/ and half of them had difficulty with /v/. She attributed this mispronunciation to the fact that the above sounds are not present in the Arabic 27 phonemic inventory and they are substituted with the voiceless and voiced counterparts /f/ and /b/. Her study was to measure the errors made by Arab learners after the age of puberty. While the current study will not be considering age, several of sounds that were investigated in the Barrow study will be a part of this study. Therefore, this study was important in providing a framework for the sounds included in this study. Messiha (1985, p. 225) stated that “it is an elementary linguistic fact” that the English phonemes which are not present in Arabic are the cause of the phonetic difficulties that Arab learners face. He added that the phonemes /p/, //, //, /v/ are good examples of the pronunciation difficulties which are mispronounced as [b], [s], [s], [f] respectively by Egyptian speakers. Barros and Messiha‟s results are important to this current study in that they provide another dimension about different performances of other Arab speakers and the types of errors they make with specific consonant sounds. 2.2 Studies focusing on general characteristics of L2 English learners Chan (2006) discussed Cantonese ESL learners‟ pronunciation of English final singleton consonants. His methodology included several different tasks; reading of a 28 word list, a description of pictures, reading of three passages and an interview. He claimed that he attempted to stimulate the learners‟ interphonology. He discovered that the participants had difficulties with the voicing contrasts of final obstruents. He also reported, among other problems, the non-release of final plosives and the inaccurate articulation of certain fricatives and sonorant consonants. The preceding vowel environments of nasals and laterals were also found to contribute significantly in determining the learners‟ acquisition of the segments concerned. This study is related to the present study because it provides good information about the coda position, since our study look into pronunciation in that region. Chan & Li (2000) conducted a detailed comparison of differences between the English and Cantonese phonological subsystems. They explored major difficulties in pronunciation that Cantonese learners may face when learning English. Cantonese does not allow voiced syllable final plosives, Cantonese learners of English tend to substitute /b, d, g/ for [p, t, k] Moreover, they transfer articulatory habits of not releasing final plosives to English. A word like step is pronounced [step˺] and not would be pronounced [not˺]. Minimal pairs such as rope and robe, maid and mate are practically indistinguishable (Chan & Li, 2000). 29 Cantonese learners exhibit the same problem of devoicing in word initial position when pronouncing voiced plosives. As a result, they tend to devoice these sounds /b, d, g/ and pronounce them incorrectly as [p, t, k]. In relation to the voiced labiodental fricative /v/, Chan & Li (2000) referred the pronunciation errors performed by Cantonese learners to the substitution of the nearest sound in their native language. The same method of devoicing is applied; as in /lv/ which will be pronounced as [lf]. Chan & Li (2000) also examined nasals, laterals and the approximant //. It appears that Cantonese speakers of English often have problems distinguishing /l/ from /n/ or / / word initially. In words like „ride‟ /d/ which Cantonese speakers pronounce it as [ld]. Some speakers may substitute /w/ for //, saying [wd] for „ride‟ /d/. The difficulties in pronunciation that Cantonese speakers of English encounter is attributed to fundamental differences between the phoneme inventories of the two languages, the characteristics and distribution of the phonemes, the permissible syllable structures and the rhythmic patterns of the two languages in question (Chan & Li, 2000). The variations in the sound systems of two languages may cause difficulties for adult second-language (L2) learners. Chinese learners of English have difficulty 30 producing English syllable final consonants and consonant clusters because few such sequences occur in the major dialects of Chinese (Flege & Wang, 1989; Kenworthy, 1987; Rogers & Dalby, 2005). However, Flege & Davidian (1984) illustrated that phonological rules may be transferred from the first language (L1) inappropriately to the L2. In a phonetic implementation comparison study of voiced stops produced in Arabic by Saudi Arabians and by both American and Saudis in English, it was found that when Saudis pronounced /b/ it was indistinguishable from /p/ to native speakers of English. However, when measuring the timing of /p/, it suggested that Saudis did realize the phonological nature of the phoneme /p/ (Flege & Port, 1981). In Flege & Port‟s study, the researchers‟ subjects were from Saudi Arabia and spoke a distinct colloquial Arabic dialect (from central and northern regions), Arabic accented English, and American English. In a similar study conducted by Flege (1980), in which he contrasted English and Saudi Arabic voiced and voiceless stops /b d g/ and /p t k/, he explained that there are differences in voiced and voiceless stops both word initially and word finally. He 31 attributed this difference to the closure interval of stops in Arabic being longer than it is in English (Flege, 1980). Although, Weinreich (1953) and Lado (1957) attributed all substitutions made by second language learners to the nonexistence of those sounds in the speaker‟s native language, it was established that not all transfer substitutions are due to the absence of that sound in the learner‟s native language. Contrary to Weinreich‟s and Lado‟s findings, after examining the production of sounds of Hungarian speakers of English, Nemser (1971) detected that some Hungarian speakers produced [] for the English // which was not a phoneme present in English or Hungarian. Although these errors were not classified as a result of native language transfer, they were categorized as developmental and were typical of what L2 learners go through in their acquisition process. Additionally, Weinberger (1990) established that the contrastive analysis adopted by Weinreich (1953), and Lado (1957) was not able to explain the different substitution errors made by Japanese and Russian speakers of English when pronouncing the English interdental fricatives // and //. In this study, Japanese speakers substituted [s] and [z] for the English // and //, respectively, while Russian speakers substituted [t] 32 and [d] for the interdentals. As a result, Weinberger embraced underspecification theory and attributed the Japanese and Russian speakers‟ error types are due to the fact that /s/ and /z/ are the least marked segments in Japanese, whereas /t/ and /d/ are the least marked segments in Russian. Vidovic (1972a), in his contrastive analysis study, compared the sound systems of Serbo-Croat and English to predict the difficulties of pronunciation of learners of both languages. In his comparison, he used words that contained the same sounds in both languages and asked native speakers, who had not yet been exposed to these words, to pronounce them. For example, he would ask a native speaker of English to pronounce the Serbo-Croat word sit /st/ which means ”full up” and he would ask native speakers of Serbo-Croat to pronounce the English word sit /st/. Then he analyzed each sound and reported the differences. Vidovic (1972b) pointed out the small differences between the two systems despite the existence of a large number of similar sounds in both systems. He registered all the differences between the Serbo-Croat and English consonants then concluded that consonants are roughly equivalent but not one is truly identical. For instance, the Serbo- Croat sound /s/ is a strong dental sound and more hissing than the English /s/ sound 33 which is an alveolar sound with less hissing. He explained that dental fricatives // and // do not exist in Serbo-Croat and they are often replaced by the sounds [t] and [d] respectively and rarely replaced by the sounds [s] and [z] noting that all four sounds are dental in Serbo-Croat. Collins & Rodd (1972) studied the pronunciation of English of West African learners from 22 different first languages. They found many sounds that caused difficulties; the labial fricatives /f/ and /v/ (in some students) were replaced by the bilabial fricatives [] and [] especially in intervocalic positions, the dental fricatives // and // were a constant source of difficulty and were replaced with the labial fricatives [f] and [v], overcorrection of /s/ and /z/ lead to the occasional replacement with [] and [], alveolar plosives /t/ and /d/ were confused with the dental [t] and [d], /p t k/ often lacked aspiration in initial position resulting in [b d g], /t d/ usually leveled with [] and // especially in medial position and /r/ is confused with [l]. Nearly all the articles evaluated regarding pronunciation among Arab speakers cite Kharma & Hajjaj (1997) as a reference. This work attempts to identify all errors in English the Arab speakers endure. It addresses a wide range of errors in grammar, syntax, discourse, idioms and vocabulary. As far as pronunciation errors go, they 34 classified mistakes into two major categories; Mistakes in Vowels and Mistakes in Consonants. More specifically, this study is interested in the mistakes in consonants. This chapter summarized a number of studies that dealt with pronunciation errors. Despite that there are many studies that investigate pronunciation problem; there are not any studies that deal with this specific dialect of Arabic. Nevertheless, the literature presented is closely related to our study in many ways and were selected in order to determine which sounds Arab speakers and other learners of English have difficulty with and which environments these difficulties are likely to be present. 35 CHAPTER 3 METHODOLOGY 3.1 Purpose of the study and research Questions The purpose of this paper is to determine the pronunciation difficulties of Saudi learners with the pronunciation of some English sounds. More importantly, the goal of this paper is to determine exactly what environments within words the errors occur. The current study investigates pronunciation of Saudi Arabian ESL learners who speak Najdi dialect. There are three main questions this study will investigate: Question one: Do Saudi ESL learners have difficulty with the sounds /p  v/? Which sound seems to pose the most difficulty? Question two: In what context do Saudi ESL learners have the most difficulty with the sounds /p  v/? Are Saudi ESL learners more accurate in their production of these sounds when they are produced in isolation (a word list) or in connected speech (a reading passage)? Do all three sounds pattern in the same way (all three sounds more accurate in isolation or all three sounds more accurate in connected speech)? 36 Question three: In what environment do Saudi ESL learners have the most difficulty with the sounds /p  v/? Are they more accurate in word-initial position or in word-final position? Do all three sounds pattern in the same way (all three sounds more accurate in word-initial position or all three sounds more accurate in word-final position)? 3.1 Research Problem As presented earlier, the current study will investigate the pronunciation difficulties of Saudi ESL learners regarding certain phonemes; the bilabial voiceless stop /p/, the labiodental voiced fricative /v/ and the alveolar approximant //. The focus of this study lies in two areas. The first one is presenting errors of pronunciation regarding Saudi Arabian Najdi dialect ESL learners of English, more specifically, which of the sounds /p/, /v/ and // cause more difficulty. The second section of this study is to see which of the problematic sounds is more difficult to pronounce and in which position. More precisely, do Saudi Arabian Najdi dialect ESL learners of English have more difficulty pronouncing the sounds /p/, /v/ and // in the onset position or coda position? Which of the three studied sounds is pronounced closer to the native sound 37 will also be analyzed in order to determine which one is the most difficult to acquire and which is acquired first. 3.2 Subjects The subjects for this study were carefully chosen to represent a certain group of Saudis. Gender and age factors were controlled and the level of proficiency was also controlled. Five subjects were chosen to participate in this study. All five subjects are from the center region of the Arabian Peninsula. Najdi dialect is the sole dialect of the region (Ingham, 1982; Ingham, 1986; Ingham, 1997; Al-Sweel, 1987; Prochazka, 1988). One of the subjects reported to be from a different city called Al-Wadi and the other subjects were from the capital city of Riyadh, nevertheless all participants spoke the targeted regional dialect of the study. The significance of the Najdi dialect among Arabic dialects within modern linguistics is the hypothesis that it is the closest dialect to Classical Arabic. In addition, it is the least studied dialect among Arabic dialects (Al- Sweel, 1987). For these reasons the subjects that spoke this specific dialect were chosen to be a part of this study. 38 3.2.2 Age, Gender and English Proficiency Age and gender were controlled to eliminate their effect on the results of the study. All subjects were males and ranged from 20-29 years of age. No subject had ever left his country of origin for any reason before coming to the US, thus they were not exposed to any other language but Arabic. Their exposure to English was relatively similar which was merely through the public educational system. All participants started learning English at the age of 12-13 in which students normally start their first year of middle school. The public education system in Saudi Arabia is divided into three levels; elementary (six years), middle school or intermediate (three years) and high school (three years). Starting from middle school to high school, students are taught English as a foreign language along with other disciplines. Students of English in Saudi Arabia are taught by Arabic native speakers and not exposed to the pronunciation of native English speakers. All participants arrived in the United States and immediately enlisted in the English as a Second Language Program in the Center of English as a Second Language (CESL) at Southern Illinois University Carbondale (SIUC). Upon their arrival to CESL, they were placed in the lower level based on their placement tests scores. Their total English exposure in the US was about 24-30 months at the time of the collection of the 39 data and they had all finished one semester in their academic undergraduate studies. None of the subjects received any pronunciation or extracurricular speaking classes other what was already provided in the ESL program. Four of the participants‟ recent TOEFL scores ranged from 460-490 on the Institutional TOEFL exam, one subject scored 527 on the same exam. 3.2.3 Educational background All participants finished an associate degree prior to their arrival in the US. After they finished their English requirements, each subject was interested in a different area of study at SIUC. At the time of data collection, subjects were studying Management, Radiology, Economy, Information System Technology and Instructional Technology. 3.3 Instruments There are two parts to the research instrument, a word list and a reading passage. Both parts of the instrument contained the same set of target words, which were selected to elicit each targeted sound (/p/, /v/ and //). The sounds were placed into 40 words in an initial position (onset) and in the final position (coda). An attempt was made to control the environment of each sound (previous or following sound). For example, for the sound /v/, the words veal /vil/ and leave /liv/ were chosen because in the first word (representing onset position) the sound following /v/ is /i/ which is the same sound as the /i/ preceding /v/ in leave. For each sound, four words were chosen in each position (initial and final) which led to a total of eight words per sound times three sounds (/p/, /v/ and //) for a total of 24 words for all three targeted sounds. The word list, referred to throughout this study as Words in Isolation (WI), included the 24 targeted words plus 12 additional distracter words for a total of 36 words to be produced in isolation (WI). The reading passage, referred to as the Words in Context (WC) instrument, included the same 24 targeted words dispersed in a reading passage which contained a total of 253 words (see appendix A and C). Upon agreeing to participate in this study, each subject was presented with a consent form and demographic information form and was explicitly asked by the researcher if he had any questions. Some participants asked for the purpose of the study and they were told that this study was looking into pronunciation errors made by Saudi 41 ESL learners. They were not told which particular sounds that the study was investigating. 3.4 Procedure The recording was completed in a quiet room on campus. Subjects were given the reading narrative words in context (WC) as their first task. They were given 5-10 minutes to read the narrative carefully and become familiar with the content then give the permission to start recording. None of the participants exhausted the allocated time reading the narrative thoroughly. They all asked to be recorded before five minutes had passed. A microphone (Labtec PC Microphone) was used for the recordings. Digital sound files were recorded immediately onto a laptop computer (Dell Inspiron 710m). Data were later analyzed and transcribed on a separate desktop computer with the help of spectrographic software (SIL Speech Tools 2.0 and Speech Analyzer Version 2.5, available from www.sil.org). For the first task which was words in context (WC), each participant read the narrative twice with normal reading pace and they were recorded both times. The second task which was words in isolation (WI) was administered 42 immediately after the reading task. The 36 words in isolation were printed on flash cards in random order and the researcher held them up for the participants to read aloud. The rationale behind this procedure was to regulate the pace of the task. If the participants had actually held the flash cards, it was expected that they would read too quickly and that might negatively affect the produced sounds. Each subject was expected to read each word aloud twice, for a total of four repetitions of each word during the recording session. 3.5 Coding After collecting the data from all participants the recordings were saved as digital sound files and labeled individually for easy access. Each subject had a main file and each file contained two main categories; Words in Isolation (WI) and Words in Context (WC). Each category contained two more subcategories; first attempt and second attempt. Since the subjects were asked to read each word in the word list twice, each word resulted in a total of four sample tokens. For consistency, only the second instance of each word pair was taking into consideration for this study. 43 Later, each sound was analyzed and the targeted sounds transcribed using the Speech Analyzer program which allowed the researcher to examine each sound via waveform and spectrogram. Every segment of a target sound the subjects produced was recorded and registered. However, only correct pronunciations were taken into consideration in this study. All data collection and transcription was performed by the researcher who is currently a graduate student of linguistics at SIU. Another current graduate student of linguistics at SIU independently analyzed and transcribed the sounds and the two transcriptions were compared. A third reviewer who was a professor of linguistics at SIU was consulted when there was disagreement in the transcription. After separating each sound and analyzing it thoroughly, each produced sound was placed in its actual position in a work sheet. Each subject was allocated four similar work sheets; first and second attempts for words in isolation (WI) and first and second attempts for words in contexts (WC) (see appendix D). An Excel database was constructed to hold all tokens produced by subjects. Although, work sheets held every pronounced utterance correct or mispronounced, this study did not take into consideration mispronunciation. Only the correct pronunciations were taken into consideration in the data base. Each correct 44 pronunciation of a sound was marked as one, while each mispronunciation was regarded as zero in the data base. 45 CHAPTER 4 RESULTS The purpose of this study was to identify and analyze pronunciation difficulties experienced by Saudi Arabian ESL speakers regarding certain English consonant phonemes. More precisely, three sounds were selected for this study; the voiceless bilabial stop /p/, the voiced labiodental fricative /v/ and the alveolar approximant //. The results of this study revealed that subjects performed differently regarding each sound in each environment. The results will be reported in the light of the proposed research questions. After counting all the correct instances of the studied sounds, it was evident that all participants faced pronunciation difficulties with each sound. These sounds are in the order of difficulty as follows: the voiced labiodental fricative /v/, the voiceless bilabial stop /p/, and the alveolar approximant //. These results coincided with Eckman‟s (1977) Markedness Differential Hypothesis (MDH). His phonological hypothesis proposed that the more common phonetic features among world‟s languages the less difficult those features are to acquire, but less common features would be more difficult to acquire. This means a more marked sound is more difficult to learn than a less marked sound. In 46 our case fricatives are more marked than stops. The current study is not concerned about the different representations of sounds the subjects produced but is concerned with the number of correct attempts that were produced. The results of this study are as follows. The total number of tokens was 240 instances. This was divided among five subjects. For every targeted sound, each subject read 16 words, 8 words in context and 8 words in isolation. Words in context and isolation were further divided into onset position and coda position which resulted in 4 tokens for each. Thus, the breakdown is 4 words times 2 positions times 2 contexts times 3 sounds times 5 subjects (4x2x2x3x5=240). Table 4.1 Total Number of Correct instances in WI and WC combined Total p v  Number Correct 136 51 34 51 Number Possible 240 80 80 80 Percent Correct 56.7% 63.75% 42.5% 63.75% 47 4.1 Question one: Do Saudi ESL learners have difficulty with the sounds /p  v/? Which sound seems to pose the most difficulty? To answer the first research question the data was analyzed collectively as shown in Table 4.1. This allowed for the production of all participants in the first and second task combined to be evaluated. When all three sounds are combined, the subjects only managed to pronounce these sounds correctly 56.7% of the time. As for the individual sounds, the percentage correct ranged from 42.5% to 63.75% with no sound being produced correctly by all subjects all the time. The literature suggested that Arab learners have difficulty with the targeted sounds. These results support that Saudi Arabian ESL learners of English do indeed have difficulties regarding pronunciation of these particular sounds as shown in Table 4.1. 4.2 Question two: In what context do Saudi ESL learners have the most difficulty with the sounds /p  v/? Are Saudi ESL learners more accurate in their production of these sounds when they are produced in isolation (a word list) or in connected speech (a reading passage)? 48 Do all three sounds pattern in the same way (all three sounds more accurate in isolation or all three sounds more accurate in connected speech)? Given the overall results in Table 4.1, it would appear that /v/ is the most difficult sound for Saudi Arabian Najdi dialect speakers to produce in English with 42.5% accuracy compared to 63.75% for both /p/ and //. However, in order to answer the second research question more thoroughly, the production accuracy by task (WI vs. WC) needs to be evaluated and the results compared according to each sound. Table 4.2 shows the results of the production task in words in isolation (WI). When examining the correct sound production in WI it is apparent that the voiced labiodental fricative /v/ was again a difficult sound to pronounce. Only 57.5%of the time subjects pronounced this sound correctly. This is followed by the alveolar approximant //, which comes second in difficulty with 80% in pronunciation accuracy. Finally the voiceless bilabial stop /p/ was the least difficult among the three with 82.5% correct pronunciations. 49 Table 4.2 Correct production in WI Total p v  Number of Correct 88 33 23 32 Number of Possible 120 40 40 40 Percent Correct 73.3% 82.5% 57.5% 80.0% The results of the word in context (WC) task are shown in Table 4.3. Subjects were less accurate in producing these sounds in context than in isolation as expected. The voiced labiodental fricative /v/ proved extremely difficult with only 27.5% correct pronunciation. However, the voiceless bilabial stop /p/comes in second in difficulty in pronunciation with only 45% of the time. Production of /p/ proved slightly more difficult for subjects in the WC than the least difficult sound in the task, the alveolar approximant // with 47.5% accuracy in pronunciation. 50 Table 4.3 Correct sounds production in WC Total p v  Number of Correct 48 18 11 19 Number of Possible 120 40 40 40 Percent Correct 40% 45% 27.5% 47.5% Table 4.3 clearly provides an answer to question 2; pronunciation of all three sounds was less accurate in context (WC) than in isolation (WI). /v/ is the most difficult sound both in context and in isolation for these Saudi Arabian ESL speakers of Najdi dialect. Our subjects also faced difficulty in the remaining two sounds, but both sounds were produced more accurately than /v/. While /ɹ/ appears to be slightly more difficult than /p/ in isolation (WI) and /p/ appears to be slightly more difficult than /ɹ/ in isolation, these differences are too close to make such a determination. 4.3 Question three: In what environment do Saudi ESL learners have the most difficulty with the sounds /p  v/? Are they more accurate in word-initial position or in word-final 51 position? Do all three sounds pattern in the same way (all three sounds more accurate in word-initial position or all three sounds more accurate in word-final position)? Table 4.4 /p/, /v/ and // production by word position Word position /p/ /v/ /ɹ/ Initial 32/40 18/40 33/40 Final 19/40 16/40 18/40 At this point a number of things have been established. After examining table 4.4, it appears that all sounds occurring in word initial position appear to be easier than those occurring in word final position. The dominance of the difficulty of the sound /v/ in both word initial and word final position is obvious, although it is less apparent in word final position than in initial position. The data suggests that the /p/ sound was more accurately produced word initially than it was word finally. Likewise, the same applies to // sound where it was produced more accurately word initially than word finally. Furthermore, // seems to be slightly easier for learners than /p/ word initially, while word finally the /p/ appears to be slightly easier than //. 52 Another dimension is added to this data by analyzing it by task as well (WI vs. WC). In the discussion of research question two, the performance of subjects in the word list was more accurate than their performance words in context. Tables 4.2 & 4.3 demonstrate that participants managed to correctly produce 73.3% of the time in the words in isolation (WI) task compared to only 40% in words in context (WC). 4.4 The voiceless bilabial stop, /p/: Unlike its other Semitic sisters such as Hebrew, Aramaic, Amharic and Socotri, Arabic lacks the voiceless bilabial stop /p/ (Newman, 2002). Some authors argue that the sound /p/ does exist in Arabic, not as a phoneme, but as an allophone of the bilabial voiced stop /b/. Newman (2002) suggested that the voiced bilabial /b/ is often devoiced in a voiceless environment. The production for the phoneme /p/ in the current study revealed an interesting contrast between its correct productions in each environment. 53 Table 4.5 Isolation and Context for /p/ p Correct % p Correct % WI initial 20 20 100 WC initial 20 12 60 final 20 13 65 final 20 6 20 Combined 40 33 82.5 Combined 40 18 45 By examining Table 4.5 above, it becomes apparent that sound production in words in isolation in both initial and final environments are superior to sound production in words in context. With both initial and final contexts combined, participants correctly produced /p/ 82.5% of the time in WI compared to 45% in WC. Moreover, subjects were able to accurately pronounce all 20 instances 100% of the time for the sound /p/ initially and 65% of the time in word final position in WI. This is in contrast to the 60% of the time that the same subjects correctly produced the sound /p/ initially and 20% of the time in word final position in WC. This noticeable distinction in word-initially in isolation is by far the highest number of errors that subjects produced in this study. 54 Although the overall performance of participants in the phoneme /p/ seemed better than the other two sounds, there are some variations on the individual level. In Figure 4.1 and Table 4.6, the individual data for the production of /p/ is totaled across the two tasks (WI and WC). After comparing the performance of individual subjects in both tasks, subject 1 managed to accurately pronounce /p/ only 43.75% of the time while subject 5 produced /p/ correctly 87.5% of the time. Figure 4.1 Individual production of /p/ Table 4.6 Subjects‟ productions of /p/ /p/ Total # S1 % S2 % S3 % S4 % S5 % WI & WC 16 7 43.75 10 62.5 8 50 12 75 14 87.5 0 10 20 30 40 50 60 70 80 90 S1 S2 S3 S4 S5 /P/ 55 4.5 The voiced labiodental fricative /v/ As established earlier, Saudi speakers have the greatest difficulty in pronouncing the /v/ sound compared to other sounds. The differences in performance between the two tasks and positions are also apparent. Table 4.7 below, displays the varying results of each subjects‟ individual production. This data demonstrates that the performance in WI was generally better than in WC, as expected. Informants produced /v/ correctly 57.5% of the time in WI compared to a low 27.5% in WC. Table 4.7 Isolation and Context for /v/ Correct % v Correct % WI initial 20 11 55 WC initial 20 7 35 final 20 12 60 final 20 4 20 Combined 40 23 57.5 Combined 40 11 27.5 The performance of subjects was slightly better word finally than it was word initially in WI. In this task, informants produced /v/ correctly 60% of the time word finally in comparison to 55% word initially. This difference gave an indication that 56 word initially, it was slightly more difficult for the subjects to pronounce /v/ than word finally. This is the opposite of the results earlier for the /p/ in the same environment. Strikingly, this was not consistent in the second task. In context, word finally appears to be more difficult than word initially for /v/. Subjects pronounced /v/ only 20% of the time correctly compared to 35% word initially in WC. On an individual basis, the subjects‟ performance was unusual and diverged from their performance in the voiceless bilabial stop /p/. Subject 3 managed to pronounce /v/ 87.5% of the time correctly, in comparison to his earlier performance in the phoneme /p/ where he fell second to last only pronouncing 50% of the time accurately. Moreover, subjects 5 and 4 who scored 87.5% and 75% respectively in their performance in the sound /p/, they scored the lowest of all participants with an equal 18.75% correct productions regarding the phoneme /v/. (See figure 4.2 & table 4.8) Table 4.8 Subjects productions of /v/ /v/ Total # S1 % S2 % S3 % S4 % S5 % WI & WC 16 8 50 6 37.5 14 87.5 3 18.75 3 18.75 57 Figure 4.2 Individual production of /v/ 4.6 The alveolar approximant // Arabic does not have /ɹ/ in its phonemic inventory. It only has the alveolar trill /r/ (Rogers, 2000; Prochazka, 1988; Al-Ani, 1970; Watson, 2002; Kharma & Hajjaj, 1997). Nevertheless, the difficulty of this sound was equal to /p/ with 63.25% accuracy. Generally, subjects‟ performance on /ɹ/ mirrored their performance in the phoneme /p/ in both tasks and positions. Informants produced [ɹ] 80% of the time correctly word initially (WI) compared to a 47.5% in WC. The data also demonstrates the difficulty of WC production with only 15% compared to the 30%. (Table 4.9). 0 10 20 30 40 50 60 70 80 90 S1 S2 S3 S4 S5 v 58 As shown in table 4.9 below, production of words in isolation in both initial and final environments is better than words in context. While participants produced [ɹ] 85% of the time word initially in isolation (WI), they only managed to produce 80% accuracy in word context (WC). Moreover, subjects were able to correctly pronounce /ɹ/ 75% of the time word finally in WI compared to a very low 15% of the time in WC. This obvious difference word finally in WC is the lowest number that subjects produced in this study. It is even lower than the production of the /v/ which was produced 20% in word final position in WC. Table 4.9 Isolation and Context for //  Correct %  Correct % WI initial 20 17 85 WC initial 20 16 80 final 20 15 75 final 20 3 15 Combined 40 32 80 Combined 40 19 47.5 On the individual level, subjects 4 and 5 scored the highest in comparison to the other participants with a 75% and 81.25%, respectively. Subjects 3 and 4 equally scored 59 62.5% in their correct performance regarding this sound where subject 1 scored the lowest with a 37.5% of the time (see figure 4.3 & table 4.10). Table 4.10 Subjects‟ productions of // // Total # S1 % S2 % S3 % S4 % S5 % WI & WC 16 6 37.5 10 62.5 10 62.5 12 75 13 81.25 Table 4.11 demonstrates the performance of each subject in each sound. Subject one expressed the lowest average of the five subjects with only 44% of accurate sound production. Despite that the data established /v/ as the most difficult sound of all, 0 10 20 30 40 50 60 70 80 90 S1 S2 S3 S4 S5 r Figure 4.3 Individual production of // 60 subject three performed better in /v/ than in /p/ or // with 67% which was the highest accuracy of all five subjects. While Subjects four and five performed the highest in the list in /p/ and /ɹ/, they did not fare as well in the overall results because of their poor performance on /v/. Table 4.11 Subjects‟ performances in each sound /p/ /v/ /ɹ/ Total S1 7 8 6 21/48 S2 10 6 10 26/48 S3 8 14 10 32/48 S4 12 3 12 27/48 S5 14 3 13 30/48 61 CHAPTER 5 DISCUSSION, IMPLICATIONS, LIMITATIONS AND CONCLUSION In this paper, more evidence has been introduced to demonstrate the difficulty of some English consonant to Saudi ESL learners. This chapter provides discussions and explanations for the results presented in previous chapters. It also sheds some light on the implications and limitations of the current study in addition to recommendations for future research. This chapter will be presented in the light of the data demonstrated in addition to the research questions. In order to answer question three the two main tasks, words in isolation (WI) and words in context (WC) are evaluated. Although the literature suggested some pronunciation problems which were predictable regarding Arab learners of English in relation to some sounds, the main objective of the present study was to find out if the bilabial voiceless stop /p/, the labiodental voiced fricative /v/ and the alveolar approximant // were a source of pronunciation problems to Saudi ESL learners. 62 5.1 Pronunciation Difficulty in /p/, /v/ and // The first research question was aimed to find out if the English bilabial voiceless stop /p/, the voiced labiodental fricative /v/ and the alveolar approximant // were problematic to pronounce for native speakers of Saudi ESL learners. The findings provided evidence that the designated sounds do pose some difficulty to the participant, regardless of the environment or position. These results support predictions made through the Contrastive Analysis Hypothesis. It suggested that the absence of sounds in a first language could result in the difficulty of acquisition of these sounds in a new language. Bayraktaroglu (1985) suggested that CAH can predict the sounds that may be difficult for learners but at the same time it failed to predict the outcome of those sounds. The data showed that there is a general agreement on the difficulty of the phoneme /v/ and also on the phonemes /p/ and //. However, there are unusual individual correspondences regarding /p and //, which will be addressed later in this chapter. In general, the difficulty encountered by Saudi ESL learners in acquiring the pronunciation of the English fricatives /v/ was evident. This finding seems to support Eckman‟s (1977) Markedness Differential Hypothesis, which states that common 63 sounds between L1 and L2 are less difficult or less marked than nonexistent sounds which are considered more difficult and more marked. However, all three sounds studied (/p/, /v/, and /ɹ/) are absent from the L1 inventory of these subjects and present in the L2 inventory. This should indicate that all three sounds will pose a difficulty for these learners which is also supported by this study. 5.2 The most difficult context To establish the most difficult context for these three sounds, each situation was separated. As introduced in chapter four, results varied when looking at sounds in isolation and in context. Although it seems that in both situations there was an agreement on the difficulty of one sound, there were slight disagreements on the difficulty of the remaining two. In word isolation and in context it was unmistakable that the voiced labiodental fricative /v/ posed the greatest difficulty in pronunciation as demonstrated by the lowest number of accurate pronunciation instances of the three sounds. On the other hand, the /p/ and the // were the second most difficult sounds to produce with the /p/ being the 64 second most difficult to pronounce in context and the // being the second most difficult to pronounce in isolation. In isolation it seemed that subjects placed // as the second most difficult sound followed relatively closely by /p/. The opposite happened in context where /p/ was pronounced correctly slightly more often than//. It is important to point out that there was not a considerable difference between these two sounds in their percentage of correct pronunciations. It is safe to assume that the two sounds are of approximately equal difficulty since the difference is so minute in each situation. 5.3 Onset or Coda Question three asked what environment do Saudi ESL learners have the most difficulty with the sounds /p  v/? Are they more accurate in word-initial position or in word-final position? Do all three sounds pattern in the same way (all three sounds more accurate in word-initial position or all three sounds more accurate in word-final position)? The results of this study indicate that /v/ is the most difficult sound in the following environments: word initially in isolation, word finally in isolation and word 65 initially in context. This is consistent with the overall data. However, Saudi speakers have the greatest difficulty pronouncing the phoneme // in the word final position in context. The above discussion established the difficulty of the phoneme /v/ as it applies to the subjects as a group. The next section will discuss the individual results. The performance of individual subjects was very different when cross- comparing isolation to context. This was also reflected when looking at the data of each environment more critically. In general, their performance in isolation (initially and finally) was superior to their performance in context (also initially and finally). This study did not consider the mispronunciations of subjects but only attempted to establish the general difficulty regarding the designated sounds regardless of their productions. However, to provide a better justification of the difficulty of the phoneme // of word final position in context, this study evaluates the mispronunciations of subjects in that position. One would think that because // is a sonorant sound it could be well absorbed when pronounced in context, where it is more likely to be in an intervocalic environment, but this was not the case for the participants of this study. As reported 66 earlier, subjects managed to pronounce /ɹ/ 15% of the time in word-final position in context accurately. That means they mispronounce that sound 85% of the time. 100% of their mispronunciations were productions of the trill [r], which is their native Arabic rhotic. A general explanation of subjects‟ poor performance in all sounds in word-final position in context could be L1 transfer. Since Arabic contains the sounds /b f r/ but lacks the /p v / (Altaha, 1995). Subjects are well aware of the sounds, especially when coming across them in isolation and word initially in context; therefore they are more careful in pronouncing them as they are pronounced in English. Subjects‟ word initial and final production in isolation was better than it was in word-final position in context especially with the sound /p/. Speakers of Arabic are extremely conscious when it comes to common /p/ vs. /b/ errors. They also have more chance to pay attention to their pronunciation when reading words in isolation in contrast to context. Another explanation for the subjects production errors may be the effect of orthography since Arabic spelling has more of a one-to-one correspondence between sound and symbol than English does (Tushyeh, 1996). That means that the majority of letters in the orthography are pronounced individually. Since subjects were reading from 67 a paper, they might become aware of the text itself. Therefore, their substitutions of the retroflex were all trills (Arabic [r]) and not misrepresentations of other sounds. On the other hand, a common misrepresentation was the devoicing of the sound /v/, subjects pronounced [f] instead of pronouncing /v/. As for /p/, subjects managed to produce the sound more accurately, they even pronounced it frequently with the correct aspiration as needed. Substitutions of /p/ were also voicing which resulted in the production of [b]. The results of this study agree with Altaha‟s (1995) study of pronunciation errors made by Saudi students learning English. It is interesting to note that despite of the different learning environments; where the first study was conducted in an EFL setting and ours in an ESL, the result was identical regarding the sounds /p/ and /v/. Moreover, the current study corresponds with Barros‟s (2003) general findings of sounds substitutions, but it does not agree with the order of difficulty that Barros came up with where she labeled /p/ as a more difficult sound than /v/. Although the data demonstrates an overall agreement on the difficulty of one sound, our results revealed a non-consistency in performance between participants individually. Subjects 4 and 5 were not consistent in their performance of the sound /v/ as they were the sounds /p /. In contrast to their performance on /p/ and /ɹ/, they scored 68 the lowest in the most difficult sound /v/. This could be explained based on the overall difficulty of the voiced labiodental fricative, which was previously discussed in relation to MDH. However, not all subjects performed as badly on the sound /v/. Subject 3, who remained average in his performance in /p/ and // scored the highest of all subjects in producing the sound /v/. To account for the unusual results regarding individual performances, especially subject 3, 4 and 5, the subjects‟ demographics were individually analyzed. Due to the small number of subjects, a hypothesis of this behavior could not be determined. It may be possible that subjects 4 and 5 were well aware of the common errors of the phonemes /p / and thus consciously performed better than other subjects on these two sounds. Moreover, they could have been unaware of the common difficulty of the sound /v/ and therefore performed the worst. The opposite could apply to subject 3 as well. 5.4 Limitations of the Study Despite the effort to control validity threats, this study suffered a number of limitations which could hinder its value. First, this study was limited to the pronunciations difficulties of the voiceless bilabial stop /p/, the voiced labiodentals 69 fricative /v/ and the retroflex // in isolation and in context in all positions. It is very likely that orthography and L1 phonology greatly contributed in this matter. Therefore, a more comparable selection of sounds seems to be more accurate to research in future studies. For example, investigating difficulties in laterals, stops or fricatives would be more valid to perform accurate comparisons. Moreover, the tasks used elicit sounds were a words list and a narrative. The validity of results could be improved by observation and recording of spontaneous speech. This would also eliminate the effects of orthography as Arabic is a one-to-one correspondence language and that may affect their pronunciation. Second, the findings of this study should not be generalized to all ESL learners other than the sample studied. This study was structured to monitor Najdi dialect Saudi Arabian ESL learners of English. Expanding the samples to cover another Saudi dialect will increase the importance of such studies. Finally, this study did not look into the mispronunciations of targeted sounds; therefore a definite explanation of the reason of their production could not be reached. Therefore, a limitation of this study is that it did not explain the exact substitutions for 70 errors. It is important to identify the source of the problem whether it is acquisition oriented or phonological. Despite the obvious limitations of this study, the findings hold a substantial number of implications for English teaching pedagogy. 5.5 Implications The findings of the present study revealed that Saudi ESL learners of English struggle with the sounds /p v /. In addition, it demonstrated the difficulty of the voiced labiodentals fricative /v/ in all environments and the difficulty of producing /ɹ/ word finally in context. These results study have implications both for students and teachers of English alike. Recognizing sources of difficulties in speaking may help students to be more cautious about their pronunciation. Careful pronunciations of problematic sounds can help communication and also aid non-native speakers to be better understood (Mettler, 1989; Derwing, 2003). ESL teachers should be aware of the different first languages and their contrastive characteristics in the classroom. This knowledge ensures better 71 understanding of problems and therefore they will be better prepared to address students‟ needs. Moreover, understanding the phonetic inventory of students‟ L1 can provide a better tool in predicting the source of mispronunciations by identifying the difficult sounds as this study revealed regarding the sounds /p v /. By acknowledging the difficult sounds, teachers can prepare materials, instructions and pronunciation exercises to better address them. Knowledge of the exact environments of difficulty with these sounds will assist teachers to target the problem at its roots. When instruction is required, teachers will be able to concentrate on specific sounds and environment and address them accordingly. The results of this study can positively benefit English teachers when dealing with Saudi ESL learners. 5.6 Suggestions for Future Research The findings of this study established the difficulties of the sounds /p v /. More specifically, it revealed the difficulty of the sound /v/ in all positions and the difficulty of /ɹ/ especially in word final position in context. This last finding suggests that further research is essential in that environment and regarding the phoneme // in general. A 72 larger more comprehensive study is vital to confirm the findings of this study and to provide more evidence to account for the ambiguities regarding some of the findings. Individual performances in this study were puzzling. Questions surrounding their pronunciation could also be resolved when a larger sample is available. This study was restricted to a specific group of participants. It would be very interesting to conduct a similar study targeting other Saudi dialects with subjects in both the U.S. and subjects in Saudi Arabia and compare the findings of both studies. This would help to establish the common difficulties of pronunciation among Arabic speakers. 5.7 Conclusion This study inspected the difficulties facing Saudi ESL learners regarding specific consonants. These sounds were the voiceless bilabial stop /p/, the voiced labiodentals fricative /v/ and the retroflex //. Previous research suggested that these sounds were problematic to Arab learners in general and this study agreed with those claims regarding Saudi speakers. 73 Furthermore, this study demonstrated not only the difficulty of sounds but where the difficulty according to word position. In conclusion, this study proved that the sample studied encounter difficulty in certain word positions. Saudi speakers of English, struggle with the sound /v/ in most word positions. They also have difficulties in pronouncing // in word final in context as //. The interlanguage of Saudi speakers contains instances of correct and incorrect production of the problematic sounds. This study further demonstrated that special attention is required when teaching the English sounds of /p,v, r/ when in the word final position. This study also demonstrates that it is important to teach these sounds in context rather than in isolation only. 74 REFERENCES Al-Ani, S. H. (1970). Phonology: An Acoustical and Physiological Investigation. Mouton: The Hague. Al-Kahtany, A. H. (1995). Dialectical Ethnographic 'Cleansing': ESL Students' Attitudes Towards Three Varieties of English. Language & Communication, 15(2), 165- 180. Al-Kahtany, A. H. (1997). The 'Problem' of Diglossia in the Arab World: An Attitudinal Study of Modern Standard Arabic and the Arabic Dialects. Al-cArabiyya, 30, 1- 30. Al-Sweel, A. I. (1987). Verbal and Nominal Forms of Najdi Arabic. Anthropological Linguistics, 29(1), 71-90. Altaha, F. M. (1995). Pronunciation Errors Made by Saudi University Students Learning English: Analysis and Remedy. Review of Applied Linguistics, 109-110, 110- 123. Anderson, S. R. (2005). How many languages are there in the world? Linguistic Society of America, Downloaded December 5. from http://www.lsadc.org/info/pdf_files/howmany.pdf Barros, A. M. (2003). Pronunciation Difficulties in the Consonant System Experienced by Arabic Speakers when Learning English after the Age of Puberty. Unpublished master‟s thesis, West Virginia University. Bayraktaroglu, S. (1985). On Sources of Pronunciation Difficulties in Foreign Language Learning. Anglo-American Studies, 5(2), 115-121. Burt, M. & Kiparsky, C. (1972). The Gooficon: A Repair Manual for English. Rowley, MA: Newbury House Publishers. 75 Cargile, A. C. & Giles, H. (1998). Language Attitudes toward Varieties of English: An American-Japanese Context. Journal of Applied Communication Research, 26(3), 338-356. Chan, A. M. & Li, D. S. (2000). English and Cantonese Phonology in Contrast: Explaining Cantonese ESL Learners' English Pronunciation Problems. Language, Culture and Curriculum, 13(1), 67-85. Chan, A. W. (2006). Cantonese ESL Learners' Pronunciation of English Final Consonants. Language, Culture & Curriculum, 19(3), 296-313. Chiswick, B. & Miller, P. (2007). The Critical Period Hypothesis for Language Learning: What the 2000 US Census Says. Institute of Government and Public Affairs, Collins, B. & Rodd, J. (1972). English pronunciation problems of francophonic West Africans. English Language Teaching Journal, 27(1), 79-88. Corder, S. P. (1967). The Significance of Learner‟s Errors. International Review of Applied Linguistics , 5, 11-25. Derwing, T. M. (2003). What Do ESL Students Say about Their Accents? Canadian Modern Language Review, 59(4), 547-566. Derwing, T. M. & Rossiter, M. J. (2002). ESL learners' perceptions of their pronunciation needs and strategies. System, 30(2), 155-166. Dulay, H. C. & Burt, M. K. (1974). Natural Sequences in Child Second Language Acquisition. Language Learning, 24(1), 37-53. Eckman, F. (1977). Markedness and the contrastive analysis hypothesis. Language Learning, 27, 315-330. Eckman, F. (1981). On the Naturalness of Interlanguage Phonological Rules. Language Learning, 31(1), 195-216 76 Edge, B. A. (1991). The Production of Word-Final Voiced Obstruents in English by L1 Speakers of Japanese and Cantonese. Studies in Second Language Acquisition, 13(3), 377-393. Fellbaum, M. L. (1996). The acquisition of voiceless stops in the interlanguage of second language learners of English and Spanish. Paper presented at Fourth International Conference on Spoken Language Processing, October 3-6, 1996, Philadelphia, PA. Retrieved May 11, 2008, from http://www.asel.udel.edu/icslp/cdrom/vol3/663/a663.pdf Flege, J. (1980). Phonetic approximation in second language acquisition. Language Learning, 30, 117-134. Flege, J. & Davidian, R. (1984). Transfer and developmental processes in adult foreign language speech production. Applied Psycholinguistics, 5, 323-347. Flege, J. & Port, R. (1981). Cross-language phonetic interference: Arabic to English. Language & Speech, 24, 125-146. Flege, J. & Wang, C. (1989). Native-language phonotactic constraints affect how well Chinese subjects perceive the word-final English /t/-/d/ contrast. Journal of Phonetics, 17, 299-315. Geranpayeh, A. (2000). The Acquisition of the English Article System by Persian Speakers. Edinburgh Working Papers in Applied Linguistics, 10, 37-51. Giles, H. (1970). Evaluation reaction to accents. Education Review, 20, 211-227. Gordon, Raymond G., Jr, (ed). (2005). Ethnologue: Languages of the World, Fifteenth edition. Dallas, Tex: SIL International. Online version: http://www.ethnologue.com/ Ingham, B. (1982). North East Arabian Dialects. London: Kegan Paul Internat. Ingham, B. (1986). Bedouin of Northern Arabia: Traditions of the Al-Dhafir. New York: Methuen Law Book Company. 77 Ingham, B. (1997). Arabian Diversions: Studies on the Dialects of Arabia. England: Ithaca. Isaksson, B. (1999). The Non-Standard First Person Singular Pronoun in the Modern Arabic Dialects. Zeitschrift für Arabische Linguistik, 37, 54-83. Johnson, J. S. & Newport, E. L. (1989). Critical Period Effects in Second Language Learning: The Influence of Maturational State on the Acquisition of English as a Second Language. Cognitive Psychology, 21(1), 60-99. Kenworthy, J. (1987). Teaching English pronunciation. Essex: Longman. Kharma, N. & Hajjaj, A. (1997). Errors in English among Arabic Speakers: Analysis and Remedy. Beirut, Lebanon: York Press. Krashen, S. D. (1973). Lateralization, language learning, and the critical period: Some new evidence. Language Learning, 23(1), 63-74. Lado, R. (1957). How to Compare Sound Systems. Ann Arbor: University of Michigan Press. Lenneberg, E. H. (1967). Biological Foundations of Language. New York: John Wiley & Sons. Major, R. C. & Faudree, M. (1996). Markedness Universals and the Acquisition of Voicing Contrasts by Korean Speakers of English. Studies in Second Language Acquisition, 18(1), 69-90. Messiha, G. Z. (1985). Some Phonological Problems That Face Arab Learners of English. Studies in African Linguistics, 9(Dec), 224-227. Mettler, S. (1989). Recognizing and Resolving ESL Problems in a Corporate Setting. Ann Arbor, MI: Eastern Michigan University. Mneimneh, H. (1997). Arabic: One Language or Many? Language International: The Magazine for Language Professionals, 9(1), 18-19. 78 Mukattash, L. (1981). Wh-Questions in English: A Problem for Arab Students. International Review of Applied Linguistics in Language Teaching, 19(4), 317- 332. Nemser, W. (1971). An Experimental Study of Phonological Interference in the English of Hungarians. Bloomington, IN: Indiana University Publications Uralic and Altaic Series. Newman, D. L. (2002). The phonetic status of Arabic within the worlds languages: the uniqueness of the lughat al-daad. Antwerp Papers in Linguistics, (100), 65-75. Retrieved May 12, 2008, from http://www.dur.ac.uk/daniel.newman/Apil2.pdf Owens, J. (2005). Pre-diaspora Arabic: Dialects, statistics and historical reconstruction. Diachronica, 22(2), 271-308. Pallier, C., Christophe, A. & Mehler, J. (1997). Language-specific listening. Trends in Cognitive Science, (1), 129-132. Prochazka, T. (1988). Saudi Arabian Dialects. London, England: Routledge, Chapman & Hall Ltd. Richards, J. C. (1971). Error Analysis and Second Language Strategies. Language Sciences, 71(17), 12-22. Rogers, C. L. & Dalby, J. (2005). Forced-Choice Analysis of Segmental Production by Chinese-Accented English Speakers. Journal of Speech, Language, and Hearing Research,48(2), 306-322. Rogers, H. (2000). The sounds of language: an introduction to phonetics Essex: Pearson Education Limited. Rosenhouse, J. (2007). Arabic phonetics in the beginning of the third millenium. Saarbrücken, 6-10(XVI), 131-134. 79 Ryan, E. B. & Bulik, C. M. (1982). Evaluations of Middle Class and Lower Class Speakers of Standard American and German-Accented English. Journal of Language and Social Psychology, 1(1), 51-61. Ryan, E. B. & Sebastian, R. J. (1980). The effects of speech style and social class background on social judgements of speakers. British Journal of Social and Clinical Psychology, 19, 229-33. Ryan, E. B., Carranza, M. A. & Moffie, R. W. (1977). Reactions toward varying degrees of accentedness in the speech of Spanish-English bilinguals. Language and Speech, 20 (3), 24-6. Scott, M. S. & Tucker, G. R. (1974). Error Analysis and English-Language Strategies of Arab Students. Language Learning, 24(1), 69-97. Selinker, L. (1972). Interlanguage. IRAL, International Review of Applied Linguistics in Language Teaching, 10(3), 209-231. Suleiman, M. F. (1993). A Study of Arab Students' Motivations and Attitudes for Learning English as a Foreign Language 54. 3. Ph.D. Dissertation, Arizona State Univ 3. Tushyeh, H. (1996). Linguistic Problems Facing Arab Learners of English. Review of Applied Linguistics, 111-112(May), 109-117. Vidovic, V. (1972a). The difficulties of English pronunciation for speakers of Serbo- Croat, and vice versa (1) English Language Teaching Journal, 26(3), 288-292. Vidovic, V. (1972b). The difficulties of English pronunciation for speakers of Serbo- Croat and vice versa (2) English Language Teaching Journal, 27(1), 88-94. Watson, J. E. (2002). The Phonology and Morphology of Arabic. Oxford: Oxford University Press. 80 Weener, P. D. (1967). The influence of dialect differences on the immediate recall of verbal messages. Ann Arbor, MI: Michigan Univ. Center for Research on Language and Language Behavior. Weeren, V. J. & Theunissen, T. (1987). Testing Pronunciation: An Application of Generalizability Theory. Language Learning, 37(1), 109-122. Weinberger, S. (1990). Minimal segment in L2 phonology. New sounds 90: proceedings of the Amsterdam symposium under acquisition of second language speech, 137- 179. Weinreich, U. (1953). Languages in contact. New York: Linguistic Circle of New York. Yoda, S. (2006). "Sifflant" and "chuintant" in the Arabic Dialect of the Jews of Gabes (South Tunisia). Zeitschrift fur arabische Linguistik, 46, 7-25. Yost, G. (1959). Syrian English. American Speech: A Quarterly of Linguistic Usage, 34(2), 109-15. Zhang, Y. & Wang, Y. (2007). Neural Plasticity in Speech Acquisition and Learning. Bilingualism: Language and Cognition, 10(2), 147-160. APPENDICES 82 APPENDIX A Words in Isolation (Word List) // /p/ /v/ word initial word final word initial word final word initial word final rain bear pain nape vase nave read fear peek keep veal leave rob car park cop value have rose more poor rope vote stove 83 APPENDIX B Words In Isolation with distracters dream hands have town left read rose more ice rob nave car bear keep small black rain rope stove leave fear loud pain child nurse veal value peek park vote poor blue corner cop vase nape 84 APPENDIX C Words In Context (reading story) She walked into the nave. Fear was all around her. The nape of her neck was frozen as ice. She managed to escape the cop but she couldn‟t out run the bear. It all started on an early morning in a small town. On that day, after she turned on the stove, Lucy took out a piece of veal and tomato paste from the refrigerator. She felt a rare pain in her back. She went to see a doctor. On her way out, she tripped and knocked down the vase in the corner. It had a great value to Lucy. She was very sad, but the pain forced her to leave soon. She put on her blue cape and took off in her car. On her way to the hospital, passing slowly by the park, she felt something in her left arm. When she looked at it her vein was black. She noticed a poor man with a rose and a rope in his hands. Then she heard a woman screaming at her son: you shouldn‟t rob our neighbors. When she arrived at the hospital, a nurse asked her to fill out more forms and read them carefully. Lucy asked for a copy to keep with her records. At the waiting room, there was a little child playing peek a boo with his mom. Then all of a sudden a loud voice shouting: vote and have a nice day! woke up Lucy from her dream to the sound of rain on the roof.) 85 APPENDIX D Work sheet # __ Word list SUB. #___ Attempt #___ // /p/ /v/ word initial word final word initial word final word initial word final _______ain bea_______ _______ain na______e ______ase na______e _______ead fea________ _______eek kee______ _______eal lea_____e _______ob ca________ ________ark co________ _______alue ha______e _______ose mo_______e ________oor ro______e _______ote sto______e 86 VITA Graduate School Southern Illinois University Turki A. Binturki Date of Birth: June 27, 1976 Department of Linguistics, SIUC Carbondale, IL 62901-4517 King Saud University Bachelor of Arts in Translation, College of Languages and Translation, May 2001 Thesis Title: Analysis of pronunciation errors of Saudi ESL learners Major Professor: Karen Baertsch