AN ANALYSIS OF PRONUNCIATION ERRORS OF PAKISTANI ESL LEARNERS AT UNIVERSITY LEVEL

This study investigates the pronunciation errors that Pakistani ESL learners make when speaking, especially entry-level university students who have just finished their intermediate education. The study can be significant in contributing to teaching and learning English as L2 in a Pakistani context. The research is also directed to explore the causes behind committing errors and the significant difference between errors made by the participants of both the selected universities. Using the purposive sampling technique, samples of 20 first-semester BS students from one private and one public university were collected for this purpose. Presentations of the students were recorded to collect the data, and the data was transcribed afterward to identify the errors. A qualitative approach, along with quantification of the data, was used to analyze the data, which was done following the descriptive analysis method. Gass and Selinker's (2008) error analysis model was used to analyze the data systematically; however, their model is simply an expansion of Corder's (1971) model for error analysis. Results revealed that the participants of Public Sector University were noticed to be committing most of the errors. Many factors are involved in committing errors, which were found in the data of this study; these errors are phonological, socioeconomic, and psychological factors. Findings suggest that the study can be helpful for pedagogical purposes as ESL teachers, course designers, and activity planners should consider the problems faced by the learners while planning for language learning development.


INTRODUCTION
Speaking is an oral activity that consists of producing systematic verbal utterances to convey a message or some meaning (Nunan, 2003).Past studies report the criteria to measure one's speaking ability is to measure its accuracy, fluency, and appropriateness.Thornbury (2005) sets a precise scale for the assessment of accuracy in terms of grammar, vocabulary, and pronunciation, whereas to measure fluency, the points taken into consideration are: Students speak clearly and without many pauses; students can quickly combine the words and phrases together to put simple sentences and complex ones, If the student is fluent, they can speak and convey the meaning in several ways using different expressions.
Error analysis is the process of examining the learning outcomes of the learner that they have achieved to develop their interlanguage system both in speaking and writing.In error analysis, the comparison is made between errors in the target language and the correct target language form.

The Concept of Error
Error is a result of some failure in a learner's performance (Corder, 1971).According to Norrish (1983), an error is a systematic deviation from the correct target language form when the learner has not learned it and relentlessly gets it wrong.Errors can be interlingual or intralingual.Krashen defines interlingual errors as the overgeneralization of a learner's native language rules to the target language when the learner cannot grasp the rules of the target language (Walz, 1982).While intralingual errors occur when learners extend the rules of the target language to the features where they cannot be applied, this might be due to the incomplete knowledge of target language (TL) rules.Learning a second language involves making and correcting errors and analyzing these errors through a systematic process.The identification and correction of errors are essential for a learner to improve their language proficiency.Corder (1967) states that understanding learners' errors is significant for teachers, researchers, and the learner itself.

Error Analysis
Erdogan (2005) states that error analysis is the activity that involves dealing with learners' performance to have an understating of the cognitive process they make use of when coding the input they draw from the target language, i.e., what is the underlying process of second language acquisition.So, from the above definitions, it can be inferred that error analysis is an activity that is executed to classify, identify and describe the errors made by the learners in both speaking and writing to have information about the common hitches' learners face in the target language performance.According to Corder (1971), this error-making strategy is used by the children acquiring their first language and the learners acquiring the second language.
The theory of error analysis was proposed by Corder (1967), and he declared that errors are not only the artifact of native language impediments.Corder (1971) has made a distinction between an error and a mistake.An error occurs due to the learner's lack of knowledge of grammar or vocabulary, and learners are not aware of these errors, so they cannot be recognized and self-corrected.On the other hand, a mistake is due to occasional factors such as carelessness, nervousness, etc.The occurrence of a mistake does not mean that the learner is incompetent in the language to be recognized and self-corrected.
Before the error analysis approach, contrastive analysis was used to identify and study the learners' errors.Still, later, it was realized that contrastive analysis only gives one reason for committing errors: mother tongue interference.This is not the only reason behind a learner's errors.So, Corder (1971) proposed another theory to study this phenomenon, which was named "error analysis." According to Corder (1971), a learner's errors provide evidence of the system of the language that he has learned at a particular point in the course.According to Corder (1973), errors have two objectives: theoretical and applied.The theoretical aim is to discuss what and how a learner learns when learning a second language.On the other hand, the used object makes the learner capable of manipulating knowledge of his vernacular language to learn the foreign language for academic purposes as it is concerned with pedagogical purposes.

Purpose of the Study
Pakistan is a multilingual country where English is used as a second language, and second language learners often make a lot of grammatical and rhetorical errors while speaking.Learning English in Pakistan is considered prestigious due to its colonial past and official status.Therefore, learning English at school, college, and university levels is of prime importance.Since learners learn English as a second language, they commit errors in the process of learning.Identification and evaluation of those errors are important to enhance the language learning and development process.The current study investigates the pronunciation errors made by Pakistani ESL learners at the university level while speaking, and it also explores the reasons behind these errors.

Research Questions
The current study seeks to address the following questions: 1. What phonological errors do the ESL students of private and public sector universities commit in their speech?
2. What is the difference between the errors made by the learners of both Universities and what are the possible reasons behind these errors?

Significance of the Study
The current study is noteworthy as it investigates the phonological errors made by Pakistani ESL learners in their speaking practices.Previously, error analysis in speaking practices has been the least attempted area, particularly in the Pakistani context.The study can be significant in many ways to identify pronunciation errors, and the sources behind these errors can contribute to teaching and learning English as L2 in a Pakistani setting; it can diminish the chances of making phonation errors.The significance of the study lies in the fact that it can inform the course designers and teachers about the kinds of errors that this specific ethnic group makes.Of course, course designers and educators are aware of the potential problems; they would be in a better position to set suitable intervention strategies and techniques which consider the student's needs for learning a target language that is based on phonology/pronunciation in the light of these errors.

REVIEW OF RELATED LITERATURE
There have been numerous studies on language competence and learners' performance errors in the past in many different contexts.Hsu (2013) reviewed the causes of grammatical errors of ESL teachers.Saad and Sawalmeh (2014) conducted a study to identify the kinds of errors that led to damaged speech among less proficient L2 Malaysian learners by following the framework of surface structure taxonomy.Sawalmeh (2013) researched the written compositions of 32 Arabic-speaking students of Saudi Arabia and classified errors by utilizing the error analysis method of Corder (1967).Alhaysony (2012) did a study to investigate the errors in terms of articles in writing samples based on the descriptive composition of 100 Saudi female EFL learners in the first year at Ha'il University Saudi Arabia and also examined the reasons behind these errors.Hojati (2013) examined errors in the oral presentations of a group of twenty advanced-level Iranian English learners by following James (2003) linguistic category classification.Elmahdi and Khan (2015) investigated pronunciation errors in English consonants 0by Saudi secondary school learners.Alkhasawneh (2014) attempted to unearth the types and frequency of errors in the written compositions of 26 non-English students of Ajloun National University in Jordan.The purpose of the study was to investigate the types of errors, their percentage, and the causes behind those errors.Saleh & Bahrami (2018) investigated error analysis in written compositions of 40 scientific articles written by the students and edited by the editing team at the Sharif University of Technology, Language, and Linguistics Center.
The results revealed that the reasons behind these errors are greatly related to L1 transfer and the similarity between Persian and English languages.
Past research in the Pakistani context Akram et al. (2017) investigated the internal and external phonological influence on language change.The study was conducted on Balochi speakers to find out the impact of their L1 on their L2.Moreover, the influence of social surroundings on their language was also studied.Data were taken from the students at Sardar Bahadur Khan Women University (SBKWU), Quetta.Using convenience sampling technique, a total of 9 students were selected for the investigation.Students were asked to utter a list of 20 English words and the utterances were recorded and transcribed afterwards.This study was conducted to find phonological variations, the impact of L1 on L2, and the influence of other socio-geographical factors on phonology.
Syed and Atta ( 2021) conducted a study to examine the difficulties that Pakistani learners face while pronouncing /v/ and /w/ sounds.Data were collected from 30 Pakistani speakers of English who migrated to England and had an input of their native accent for at least 5-6 hours daily.The findings revealed that there was not much improvement even after an ample amount of interaction with the native speakers.Therefore, the results showed that if the adult speakers initially adopt a dialect that does not differentiate these two consonants, they won't be able to improve it in later life as well.
Bilal et al. ( 2021) examined the influence of Urdu transliteration on the pronunciation of English words.Using random sampling technique, 10 secondary level students from the district of Rahim Yar Khan were selected for the study.Students were asked to read a wordlist of English and their pronunciations were recorded.The results demonstrated that the deviant pronunciation is caused by the influence of Urdu transliteration and problems in teaching techniques.Ahmad et al. (2011) seeks to analyze the articles and preposition errors of grade 8 students from Khyber Pakhtunkhwa.A hundred students were randomly selected, and data was collected from the written compositions of those 100 students.Jabeen et al. (2015) carried out research to look into the mistakes that students make when using articles.They chose a sample of 60 graduate students for their project, and they used written essays to get the data.Results showed that students who speak Urdu as their mother tongue had more trouble using articles than students who speak a language with an article system.Indefinite articles were harder for students to use; therefore, they left them out of their writing.Mushtaq et al. (2019) investigated the written compositions of intermediate students.For this purpose, they developed a corpus of essays written by 70 students from central Punjab.
13 codes were developed to analyze the data, and the results revealed that learners committed most of the errors in spelling and the wrong use of verbs was the second most frequently occurring error.

METHODOLOGY
The current study followed the qualitative research paradigm for the data analysis after categorizing and quantifying errors.Errors are classified according to the linguistic category classification proposed by James (1998).Gass and Selinker's (2008) model for error analysis has been applied, and the rationale behind using this model instead of Corder's (1971) framework is that Gass and Selinker's (2008) model requires both description and quantification of errors, unlike Corder's model.In Corder's (1971) framework, only identification and explanation of errors are requisite.Gass and Selinker's (2008) model is merely an augmentation of Corder's (1971) framework of error analysis in which Gass and Selinker (2008) incorporated quantification of errors and remediation steps as shown in Figure 1.

Remediate
A qualitative approach, along with quantification of the data, has been used to analyze the data, which was done following the descriptive analysis method.Students have been provided with certain topics to prepare presentations 10 minutes before their performance.Students had to speak on those topics for a minimum of 3 minutes each.Those merely reading from paper are not considered the sample for the study since some of them made notes and wrote their opinions on paper and were only reading those written materials.Purposive sampling is used for selecting the samples.For descriptive analysis of the data, certain categories of errors are determined by visualizing the data; patterns emerge, and errors are classified.Tables are added to individually present the categories with their related examples, i.e., what errors are made by a particular participant, while figures and graphs are included to provide the frequency rates of these errors.
A bar chart has also been subsumed to compare the frequency of errors committed by both the universities.
International Phonetic Alphabet (IPA) symbols have been used for the authentic phonetic transcription of the utterances to analyze the pronunciation.Students who showed deviation from the standard pronunciation of the UK and USA are taken as errors.These varieties have been taken into consideration to benchmark the standard because, in Pakistan, students are mostly exposed to only these two varieties.Moreover, Standard grammatical rules are followed to provide the correct forms of utterances.The transcriptions were thoroughly checked twice to ensure that the actual utterances were not altered in the transcription process.Informal interviews of the language teachers and the participants of this study were conducted.Teachers were asked certain questions about the kinds of errors their students usually make and the possible reasons behind these errors.The interviews were unstructured, and the researcher chose to investigate a couple of the teachers.What difficulties do the students have when speaking in their second language?What psychological factors can be at play when they make mistakes, etc.? Students were interviewed to have information on their demographics.
The population of the study was the students in the first semester from a private sector university (U1) and a public sector university (U2) who had finished their intermediate level and entered a higher level of education.The participants were all Urdu-speaking learners who had been studying English as a compulsory subject within the school curriculum since their first grade.The complete numerical data is presented in Table 1.Data collected in words 10 X 500 5000 words Data collected in words 10 X 400 4000 Total time in minutes: (10 X 5= 50) + (10 X 5= 50) = 100 minutes Total data in words: 5000+ 4000 = 9000

Ethical Considerations
Participants were informed about the research purpose before recording their presentation.The researcher tried to minimize the stress level of the participants by mentioning that this activity would not be graded.The names of the students have been concealed to ensure the privacy of the participants.All the samples have been encoded with numbers and letters to keep the confidentiality of the participants.Their names or any other information about them that might reveal their identification will not be disclosed.

Limitations of the Study
For this study, the data was collected from only two universities by considering a small sample size of twenty participants from each university.Moreover, participants' level of language proficiency varies, and it might be challenging to control this factor as participants are from different language and educational backgrounds.Due to sample size, population, and variability factors, the present study results might not be generalized to a wide variety of learners.

Errors of Pronunciation
Difficulty in pronunciation seems challenging for most L2 students and to examine their difficulty level, the speeches of the students of the first semesters from U1 and U2 were recorded.The focus of the research regarding pronunciation is on vowels and consonants both, as mentioned earlier.The data was recorded and then later transcribed.The words were phonetically transcribed using IPA symbols, and the pronunciation deviations were thus noted.Errors identified from the data depict the degree of the problem students of U1 and U2 face regarding pronunciation.The difference between errors of students of U1 and U2 was debated and investigated subsequently to have an accurate understanding of these issues and to answer our research questions.Errors in pronunciation are the first noticeable form that is observed very frequently.Generally, consonants and vowel sounds are key features to be observed in speaking.Tables 2  and 3 give shreds of evidence of this pattern of error which frequently occurs in the speaking sample of the participants taken from both the Universities, U1 and U2, as the instances corroborate the findings., and S5 did not commit any vivid mistakes in terms of pronunciation, while S3 did not commit errors regarding vowel sounds.However, S3 made a consonant sound error where the participant pronounced sentence /senta:ns/ instead of stance /stɑːns/.S4 made two vowel errors and zero consonant errors.The participant said /a:ver/ instead of our /aʊə(r)/ and /əθrized/ instead of authorized /ɔːθəraɪzd/.S6 made two vowels and one consonant error, e.g., the speaker was supposed to say entertainment /entəˈteɪnm(ə)nt/ and robot /rəʊbɒt/ but he uttered /intərteɪnmənt/ and /ri:bɒt/ respectively and for consonant instead of saying though /ðəʊ/ he spoke Thu /θu/.S7 made one vowel and three consonant errors.S7 said /raɪd/ instead of Rid /rɪd/ where they have completely changed the midvowel sound and made it a diphthong which was originally a monophthong in its native British form.
Figure 2. Frequency rate of kinds of pronunciation errors by the participants of U1.
Figure 2 indicates that vowel errors comprise 71% of the total pronunciation errors of U1 students, while they are 29% concerning consonants.Table 3 contains illustrative examples of errors obtained from U2 recordings and described along with them.
Table 3 shows that U2 students committed more mistakes in vowel sounds rather than in consonants, as most of the errors were due to the addition of a schwa (ə) sound.Each of the ten samples made vowel errors, and the examples are given below: S1 was supposed to speak "of" with the back open vowel /ɒ/ as /ɒv/, changing the consonant sound from /f/ to /v/.However, the participant incorrectly pronounced it as /a:f/, which made it an error.The addition of Schwa (ə) caused most of the errors, e.g., /raʊndə/ /bʌtə/, /ðɛnə/ and /əweɪtɪŋ/ instead of saying round /raʊnd/, but /bʌt/, then /ðɛn/ and waiting /əweɪtɪŋ/ respectively.Many of the errors occurred since participants were frequently using an open long vowel /a:/ instead of the back open vowel /ɒ/.For example, knowledge /nɒlɪʤ/ and college /kɒlɪʤ/ were observed to be pronounced as /na:lɪʤ/ and /ka: lɪʤ/ in the language of S2 and S10, respectively.More examples of vowel errors can be seen in Table 3.

Causes of Pronunciation Errors
To explore the reasons behind the errors, some of the language teachers were interviewed.From U1 and U2, five teachers, each from a total of ten, were chosen.They were subjected to a semi-structured interview in which questions centered on the most typical errors they observe students make in their speech and potential explanations for why they do so.Responses reported that there are a lot of reasons why these errors take place.Most of the respondents answered that students make many pronunciation errors, which might involve the mother tongue influence factor as they make errors due to the interference of their first language.It can be perceived from examples that participants made redundant use of "schwa" /ə/ sound even where it should have been a consonant cluster; learners put a schwa sound between the consonants.Most of the vowel errors occurred due to the excessive use of the schwa sound.Moreover, the researcher took the demographics of the students to show how many students are from urban areas and how many are from rural areas.Data delineates that U1 participants are primarily from urban areas as compared to the students of U2.Generally, the mother tongue of the students is Urdu and Punjabi.There are many English words that Urdu speakers use in their daily lives.Urdu and Punjabi speakers pronounce English words with the influence of their phonological system.According to past researchers, borrowed words always undergo a phonological change influenced by the recipient language.Therefore, many pronunciation errors occur due to this fact.Furthermore, many of the respondents mentioned that psychological factors are also involved in making pronunciation errors.When the speaker feels nervous and has anxiety about the audience, they get the words wrongly pronounced due to this overwhelming fear and anxiety.

CONCLUSIONS AND IMPLICATIONS
The study was designed to examine the pronunciation errors Pakistani ESL learners make while speaking and how frequently these errors occur.To answer this question, the sample was selected from two reputed universities: a public and a private sector university.As discussed in the methodology section, samples were selected, and data were collected using the convenience sampling technique.Pronunciation errors were identified by visualizing the data and International Phonetic Alphabets (IPA) were used to transcribe to analyze the data.The analysis has also attempted to explore the causes behind such errors, and for this purpose, informal interviews were conducted with some of the language teachers.It was reported that there are various phonological and psychological reasons why students commit errors.The study also indicates the difference between the errors by the learners of the BS level of U2 and U1.U2 participants committed numerous errors resulting from the influence of the phonological system of their mother tongue.As mentioned in the causes, most of the students have come from rural areas, and in Punjab, people who live in villages usually speak Punjabi.Punjabi and Urdu do not allow consonant clusters; therefore, participants frequently added a Schwa /ə/ sound while speaking the target language due to this factor.Although these errors cannot be considered as mistakes, they become non-standard.As far as socioeconomic and psychological reasons are concerned, both the socioeconomic and psychological Vowels 77% Consonants 23% Pronunciation Errors Vowels Consonants components have influenced the learning outcomes.Students of U2 have come mainly from rural areas where people are usually uneducated or less educated with average social class status.U2 is a government institute that provides education at low rates compared to private universities, and it is situated in a countryside area.Therefore, students from the surrounding rural areas find it accessible economically.Most of the students have completed their matric and intermediate from Government schools and colleges.
In Pakistan, the condition of these schools and colleges is very abysmal.Poor teaching conditions led the students to lose confidence.In terms of language learning, learners do not feel confident enough to speak proficiently or even express themselves intelligibly to the audience.This incompetency creates trouble for them when they enter the university level.
The findings of this research yield implications in terms of pedagogy as they are significant for ESL teachers, course designers, activity planners, etc. Language teachers gaining post-graduate or post-experience qualifications in applied linguistics and English language teaching can use the corpus to deepen their understanding of second language learners' language (English) errors, eventually leading them to a satisfactory solution.The study can help language teachers identify the specific language problems of Urdu native speakers or other regional languages, how these L1s influence one another in the Pakistani framework, and how to maintain them.Based on the findings, ESL teachers and course designers can develop and design activities accordingly.As the causes of errors are also mentioned, it can help language teachers and course developers overcome the challenges second language learners face.

Figure 1 .
Figure1.Gass and Selinker's (2008) Model for Error Analysis.The recorder was used to record the presentations of the participants inside the classroom.All the samples were encoded with numbers or letters in the Appendix to keep the confidentiality of the participants.

Figure 3 .
Figure 3. Frequency rates of kinds of pronunciation errors by the participants of U2.

Table 1 .
Sample and population.

Table 2 .
Examples related to pronunciation errors made by the participants of U1.