Pronouncing Inert Letters: What can we learn from undergraduate EFL students?

English Education students are prepared to be English teachers as one of the sources in learning activity. However, the students still mispronounced some words including ones which contain inert letters. This study aimed to find out the most frequent errors made by the English Education Students in pronouncing inert letters and the factors of the inert letters’ pronunciation errors. This study was a descriptive study which applied quantitative and qualitative approaches. The subjects of this study were 25 students of the English Education Study Program at Universitas Bengkulu. This study used recording, observation checklist, and interview as the instruments. Recording and observation checklists were used to gather subjects’ pronunciation. In order to find the factors of the pronunciation errors, an interview was conducted. The pronunciation was analysed by using pronunciation checkers, Voice Notebook and Audacity. The interview results were analysed based on factors of pronunciation accuracy. The results showed that the most frequent errors of inert letters pronunciation made by the subjects were sounded inert letter /d/ of /-nd/ by 94% and silent inert letter /a/ of /-cal-/ by 88%. The interview generated some data which shows that the factors of subjects’ pronunciation errors were native language, age, and amount of exposure to English.


Introduction
Indonesians speak English infrequently in daily conversation since English is as their foreign language. It is very rare to hear people speaking English at public places in this country. Most of them usually speak in their mother tongue. It can be Bahasa Indonesia or a vernacular language based on which region they are from. In addition, spoken English in schools is used only in English courses. It indicates that there is only very little chance for Indonesians to practice speaking English for communication even for English Education students.
However, there are many components to notice in speaking. One of them is pronunciation. In order to be fully understood in a conversation, people should produce the words with correct pronunciation. Pronunciation in English is quietly different from pronunciation in Bahasa Indonesia. Most of the words in Bahasa Indonesia are pronounced as how exactly the words written while pronunciation in English is not. It makes pronunciation in English becomes more confusing and varied.
This term of pronunciation is learned in one of subjects in the English Education Study Program at Universitas Bengkulu. It is called phonology. By its definition, phonology is a branch of micro-linguistics that studies how to organize sounds. Studying sounds organization in English is more complex since there are so many rules within the sound. Pallawa (2013) stated that based on his research, Indonesians have difficulty in pronouncing English words since there are many phonological rules in English which do not exist in Bahasa Indonesia. That may be one of reasons why it is very often to hear many English students still mispronounce some English words.
The big numbers of incorrect pronunciation is proved on some previous studies which are done in the same place as this research will be conducted, the English Education Study Program of Universitas Bengkulu. The finding indicates that there are still a lot of English Education students who pronounce some English words incorrectly. The first one is a study conducted by Irianto et al. (2018) which examined the pronunciation errors of English Consonants /θ/ and /ð/ by the English Education Students of the Universitas Bengkulu. The result showed that most of the students pronounced those two consonants /θ/ and /ð/ incorrectly. The reason is both of the consonants do not appear in Bahasa Indonesia so that they pronounced them like they pronounced the consonants /t/ and /d/ in Bahasa Indonesia. was conducted by Styvant et al. (2019). His research is about how the students of English Education Study Program at Universitas Bengkulu pronounce English syllable words stress. The research shows that English students have average predicate in pronouncing syllable words stress in the first and second syllable and low predicate in two stresses in a word.
There are some previous studies related to this research done in Yogyakarta. The first one was conducted by Utama (2018). It focused on silent letters "H" and "W" pronunciation by Students who were not in English Department. The second one was by Ayumi (2018). The study focused on silent letter "G" pronunciation by elementary and advanced level students at English Made Easy (EME) course. There is another previous study conducted in Saudi Arabia by Alqunayeer (2016). The study focused on letter "G" pronunciation made by Saudi Undergraduate Students. The result showed that the most frequent errors occurred in the letter "g" before nasal which could be categorized as silent letter. However, this research investigated silent letter in different focus; that is inert letter pronunciation by Undergraduate English Education students.
This research mattered to be conducted since this study which examined this type of silent letter is still rare. There are some previous studies related to silent letter but the focuses are different from this research. In addition, English Education students are prepared to be English teachers that they must comprehend the language itself, including inert letter pronunciation. Those are the reasons why the researcher would like to investigate Undergraduate English Education students at Universitas Bengkulu's pronunciation on inert letter and the factors of the error in research under the title "An Analysis of Pronouncing Inert letters by Undergraduate English Education Students at Universitas Bengkulu." This research was expected to be an answer whether the students are ready or not to be English teachers in terms of inert letter pronunciation. Document strictlt." (Smith, 2020, p. 5).

Phonology
Phonology is a branch of micro-linguistic that studies about human speech sound in a particular language. Another study that is related to speech sound is phonetics. McMahon (2002) states that phonology is as narrowed-down phonetics. For instance, babies who are in babbling phase, produce some sound which they have heard in their environment as they imitate how people produce words. This process is called phonetic. The babies then narrow down the sound that they received into a certain language that they hearing and learning. This one is called phonology. Odden (2005) stated that phonology is a study of sound structure in a certain language. Phonology concerns with the implementation the sound in a certain language to real conversation. According to Davenport and Hannahs (2005), phonology deals with how human speech sounds are organized into system in a particular language, for example; how the sounds combine and how they affect to each other.

Consonant Cluster
Consonant cluster is one of syllable features in English phonology. Roach (2009) explained that consonant cluster is when a word has two or more consonant together. He mentioned that there are 2 types of consonant cluster according to its position; initial and final consonant cluster. In addition, Kreidler (2004) mentioned that consonant cluster also has another position in a word. It is medial cluster, such as the cluster /-ngl-/ in English.

Omission and Insertion of a Consonant
There are a lot of consonant clusters in English. It contributes the fact that some spelling in consonant clusters is omitted and inserted. Kreidler (2004) stated that the clusters are often simplified which makes the spelling becomes omitted, as in kindly that has the cluster /-ndl-/ or windmill that has the cluster /-ndm-/.

Morphophonemic
Morphophonemic is a study which morpheme and phoneme affects each other. According to Dwivedi (2017), morphophonemic is a system where sound structure, known as phonology, and word structure, known as morphology, interact and affect each other. Because of the interaction, pronunciation of a morpheme may get modified or changed. Morphophonemic process, according to Nurhayati (2015), is the process of phonological realization of allomorphs of a particular language or the study of phonemic representation of morphemes in different condition Ampa (2019) stated that morphophonemic process is phonological evidence which happens because of morpheme encounters with morpheme.

Pronunciation
Pronunciation is the way of someone to pronounce or to speak out words that is accepted or generally understood in a particular language. Every language has a different set of segments so speakers who learn a certain language have to learn its pronunciation. Diah and Rarasteja (2006) also said that pronunciation deals with the reception and production of speech sound to achieve the meaning. Different pronunciation may achieve different meaning. Nunan (2015) stated that speaking requires someone to be competent in articulating the sound of the language, having appropriate vocabulary, and mastering the grammar component.

Aspects of Pronunciation
There are seven aspects in pronunciation suggested by Kenworthy (1987), they are; combinations of sounds, linkage of sounds, word stress, rhythm, weak forms, sentence stress, and intonation.

Factors of Pronunciation Accuracy
The most ideal of pronunciation is native-like pronunciation. But in reality, there are still some errors in pronouncing English words and caused by some factors. There are six factors of pronunciation accuracy by Kenworthy (1987); the native language, the age factor, amount of exposure, phonetic ability, attitude and identity, and motivation and concern for good pronunciation. These factors were used to make questions for interview in this research.

Silent Letter
Silent letter is a letter that is written but is not sounded in the pronunciation of a word. Podhaizer (1998) stated that silent letter is present when someone is spelling the word but it disappears when the word is spoken. For example, the word "column"; the 'n' is silent when pronouncing the word but it cannot be omitted from the spelling. If the letter "n" is omitted from the spelling of the word "column" which becomes "colum", it has no meaning or it becomes a non-word. According to Carney (1994), there are three types of silent letter; auxiliary letter, inert letter, and empty letter.

Auxiliary Letter
Auxiliary letter is silent letter that makes significant different meaning to another similar word. Carney (1994) wrote that auxiliary letter is an additional letter to make a distinct complex graphical unit. This type of letter happens in the words "hour" and "our". The letter [h] makes difference in both of words although [h] in the "hour" is not pronounced.

Inert Letter
Inert letter is one type of silent letter which is actually sounded in a word but it is silent in another form of the word. This type of silent letter becomes the researcher's focus in this research. Carney (1994) pointed out that inert letter is a letter that is pronounced in a certain form of word but unpronounced in another form. The example can be seen in the letter [g] which is silent in "sign" and "signer" but is pronounced in "signature" and "signatory". This also happens in the letter [n] which is unpronounced in "damn", "column", and "condemn" but is sounded in "damnation", "columnist", and "condemnation". Clearer distinction of the examples can be seen in this following table completed by each transcription based on Oxford Advanced Learner's Dictionary.

Empty Letter
Empty letter is a type of silent letter that does not have distinctive function like Auxiliary letter and inert letter. It is just simply unpronounced. Carney (1994) classified empty letter into the consonant cluster of a word and it can only be found in the onset or the coda of a syllable. Empty letter that is in the onset happens to the letter [h] in the word "honest" and "heiress" while empty letter that is in the coda occurs in the letter [b] in the word "plumb", "thumb", and "tomb".

Subjects
The subjects of this research were the sixth semester English Education Students at Universitas Bengkulu who had learned speaking and phonetics and phonology subjects. The total students were 96. The researcher took samples for both research questions by about 25% of the total students since the minimum sample for qualitative study by Gay and Diehl (1992) is 10% of the total population. Thus, there were 25 undergraduate English Education students who were examined and interviewed for this research.

Instruments
The researcher used recording as the instrument. The students' pronunciations were recorded using default video recording application on each student's smartphone. The students were asked to record their pronunciation themselves and send the recording to the researcher via WhatsApp. In order to get the data from the right person, the students did not only record their voice but also their visual. They also introduced themselves in the beginning of the video. Then the recordings were put into an application called Voice Notebook and Audacity to check the correct and incorrect of the pronunciation.
After getting the recording from the students, the result of the students' inert letter pronunciation was displayed in an observation checklist in order to make the correct and incorrect of the pronunciation easier to be checked and calculated. The table contained 20 pairs of words containing inert letter. Most words were taken from Carney (1994) and other sources from internet based on how common the words are for Indonesian people. Each pair included a word with silent inert letter and a word with sounded inert letter. Then the researcher gave a check into the correct column if the pronunciation was right or incorrect column if the pronunciation was wrong based on the analysis.
The third instrument was interview. The researcher interviewed the subjects in order to answer the second research question, which factors that cause the pronunciation errors. The interview was conducted structurally since the researcher adapted the questions based on factors that affect pronunciation accuracy by Kenworthy (1987). Yet the researcher did not include phonetic ability because it required aptitude test to measure the ability. The interview was conducted via Whatsapp video call with the subjects. The questions asked to the students were in Bahasa Indonesia so the students were easier to answer the questions.

Data Analysis Procedures
After collecting the data, the researcher analyzed the data quantitatively and qualitatively. Wolcot (1994) stated that analysis is a term that includes three basis categories in analyzing data; they are description, analysis, and interpretation.

Recording
The researcher used two tools to analyze the primary data which was inert letter pronunciation. The researcher checked the students' pronunciation whether it was correct or not using Voice Notebook. Then, the incorrect pronunciation was analyzed using Audacity to check the inert letter pronunciation by comparing the students' pronunciation to native speaker's pronunciation. There are procedures that the researcher conducted to analyze the data: a. Checking students' pronunciation using Voice Notebook to judge whether the pronunciation is correct or not.
b. Recording American native speaker's inert letter pronunciation from Online Oxford Advanced Learner's Dictionary.
c. Doing input the students' incorrect pronunciation and the audio from native speaker into Audacity to check the inert letter pronunciation.
d. Comparing both of the audio from native speaker with the data recorded.
e. Completing the observation checklist to recap the correct and incorrect of the pronunciation.
f. Calculating the percentage of the data by using this formula:

P = 100%
P : Percentage F : Frequency of students' pronunciation in wrong way N : Number of students g. Classifying the data into a classification by Sugiyono (2008) as follows:

Interview
In analysing the interview data, the researcher used some steps as follow: a. Gathering all the interview data from recorded Whatsapp video calls.
b. Listening to the interview and writing the transcription of the interview.
c. Translating the transcription into English.
c. Categorizing the data of the interview based on the factors of pronunciation accuracy by Kenworthy (1987).
d. Describing the data into qualitative narration.

The Most Frequent Errors Made by the Students of the English Education Study Program in Pronouncing Inert Letters
The result shows that the most frequent errors are sounded inert letter /d/ of /nd/ pronunciation by 94% of its total pronunciation. It is followed by silent inert letter /a/ of /-cal-/ pronunciation by 88% of its total pronunciation. The errors made by English Education students considered as very high error based on Sugiyono (2008) since both of the percentages are more than 85%. Meanwhile, the other inert letter pronunciations are considered as very low, low, and moderate error since they are less than 68%. From the table, the total percentage of all the errors is about 36.7%. There were 367 errors of 1000 total pronunciations. The error number of silent inert letters is bigger than the sounded ones. The percentage of the silent inert letter is 37.6% for which there were 186 errors of 500 total pronunciations. Meanwhile, the error percentage of the sounded inert letter pronunciations is 35.8% for which there were 179 errors of 500 total pronunciations for the sounded one. The total percentage of the inert letter errors occurred is predicated as low error. From the table 4.2, it can be seen that the number of errors occurred in words containing inert letter /g/ of /-gn/ and /-gn-/ is varied from 20% to 72%. All the errors are considered as low error until high error since the percentages are less than 84%. The word with silent inert letter /g/ of /-gn/ which has the biggest percentage of words is the reign by 52%. The biggest percentage for words with sounded inert letter /g/ of /-gn-/ occurred to designation by 72%. The error of designation can be seen in the following figure.

Figure 1: Designation Pronunciation Error
The figure consists of student's pronunciation on the upper one and native speaker's pronunciation on the below one. From the figure above, the student incorrectly pronounced the word since the student did not sound the /g/ letter meanwhile it had to be pronounced like in the below part of the figure. The lines are almost flat because when someone pronounced g, like in the word designation, the vibration and the sound are blocked so there is no wave captured. The table and chart above show that there was no incorrect pronunciation for words with sounded inert letter /a/ of /-cal/, they are economical, theoretical, practical, and chemical. Meanwhile, there were many errors in pronouncing words with silent inert letter /a/ of /-cal-/. The percentages of all the words with silent inert letter /a/ of /-cal-/ are predicated as very high error except for the word economically which the percentage is predicated as high error. The biggest percentage is theoretically by 96% errors of total pronunciations. The pronunciation error of the word theoretically is shown in the figure 2.

Figure 2: Theoretically Pronunciation Error
From the figure above, the student incorrectly pronounced one of words containing silent inert letter /a/ of /-cal-/, theoretically, since the student sounded the /a/ letter clearly meanwhile it had to be unpronounced like in the below part of the figure. In the below part, the wave is small since the letter /a/ was not pronounced so that there was not much vibration. Unlike the upper part of the figure, the wave is bigger because the letter /a/ was pronounced. The table above shows that all the subjects pronounced words with silent inert letter /d/ of /-nd-/ like grandma, handkerchief, and sandwich correctly. Meanwhile, most students mispronounced words with sounded inert letter /d/ of /-nd/ like grand, hand, and sand and predicated as very high error. 24 of 25 students mispronounced the inert letter /d/ of the words hand and sand as shown in Figure  3.

Figure 3: Grand Pronunciation Error
The figure above shows that the student did not pronounce the letter /d/ in the word grand whereas the letter /d/ had to be sounded as the below part of the figure shows. At the end of the native speaker's pronunciation, there is another small wave which indicates that the letter was sounded so that the vibration existed again. The data show that there were few pronunciation errors for words with inert letter /n/ of /-mn/ and /-mn-/. There were only 0 to 6 mispronunciations of each word. The errors were occurred in the words like condemn, damn, columnist, condemnation, and damnation. The students incorrectly pronounced one of the words as shown as the figure below.

Figure 4: Condemnation Pronunciation Error
The figure shows that the student did not pronounce the letter /n/ of the word condemnation. It can be seen from the marked decrease wave which indicates that there was no the sound of /n/ after /m/. Meanwhile the native speaker's pronunciation wave shows that the wave is more static than the students. It means the sound of /n/ exists. The table and the charts show that the most pronunciation errors happened on the word with silent inert letter /g/ of /-gm/, paradigm by 64%. There were 16 students who pronounced the word incorrectly. The error is predicated as moderate error while the other errors are predicated as low error for in the interval of 36% -51%. One of the errors occurred in pronouncing the word phlegmatic as shown in this following figure.

Figure 5: Phlegmatic Pronunciation Error
The figure above shows that the students pronounce the word incorrectly. It can be seen from the marked waves. The native speaker's marked wave seems to be flat so that the vibration was blocked. It indicates that there was the sound of /g/ in the pronunciation whereas the student's pronunciation has a bigger wave which means the vibration still existed. It indicates that the student did not sound the letter /g/ in phlegmatic.

Factors of Students' Pronunciation Errors
The interview was done to find the factors of the pronunciation error in the terms of inert letter. It involved the subjects' native language, first age learning English, amount of exposure, attitude and identity, and motivation for good pronunciation. It was conducted by Whatsapp video call. The interview was conducted from July 1 st until July 5 th 2020 where the researcher interviewed 5 subjects per day for five consecutive days. The following data were collected from 25 students as the subjects of this research. The students' interview responses were categorized into five factors of pronunciation accuracy by Kenworthy (1987) as follows:

The Native Language
All English Education students as the subjects of this research answered that their mother tongues were not English. Mostly, their mother tongues were Bengkulu language. Other students' mother tongues were Bahasa Indonesia, Serawai, Lembak, Javanese, Pasma, and Rejang language. According to the interview, Serawai language is the only one language which has inert letters as in English. The example stated by the subjects was the letter /r/ in words rumah and perumahan. The letter /r/ was not pronounced in the word rumah but was sounded in the word perumahan. It showed that the letter /r/ in the words were inert letter. Meanwhile the other students stated that all letters are pronounced as how they are written in words so there was not a rule like inert letter in their native languages. In brief, their native language, which was not English, affected to their inert letter pronunciation errors.

The Age Factor
All the students were under 15 years old when they started learning English. Most students have learned English since they were 6 to 9 years old in elementary school. There was only one student saying that before joining school, she had already begun learning English from her father at 3 years old. The others started learning English when they were in junior high school. They had learned the Basic English in academic institution such as school and courses before learning English intensively in English Education study program. It could be concluded that age affected to the students' pronunciation errors as only one student who started learning English in her golden age, 0-5 years old.

Amount of Exposure
The subjects had much exposure to English in academic environment since they were majoring English Education, yet they had lack of exposure to English in their daily life. There was only one student who claimed that she was often exposed to English in her daily life because she was surrounded by people who supported her to speak English. Meanwhile the other 24 students said that they rarely found English in their daily life. They might find English in a written form like in social media but it was very seldom for them to be involved in a daily conversation using English. Some students said that they were not supported by their surroundings to speak English. People tended to speak their vernacular language for communication in their daily lives. Hence, their lack of exposure to English gave impact to their inert letter pronunciation errors.
All the students answered that their identity as English Education students affected them to try to speak English like natives. More than half of the subjects explained that they had a responsibility to pronounce words in English as similar as native speakers did. Due to the identity, they would feel ashamed if their pronunciations were bad. Some other students said that they were motivated by the identity to be better in speaking English. Being surrounded by lecturers and classmates who frequently spoke English affected them speak English like native speakers. Some of them also said that they would like to show up their capability to speak English well in front of people since their identity were as an English Education study program student. Hence, identity and attitude seem to not be considered as a factor of the students' pronunciation errors since the students' attitude towards their identities as English Education students were positive.

Motivation and Concern for Good Pronunciation
All subjects stated they had big motivation to improve English pronunciation. When the students were asked about their motivation in scale one to ten, theirs was on at 7 until 10. The students have done several activities to improve their pronunciations. The first one was searching the pronunciation of words they did not know in dictionary or in google translate, and then they tried to pronounce them. The second one was listening to western songs and movies before trying to imitate the pronunciations. The last one was the students used social media to find native speakers talking and teaching English through videos mostly on Instagram and YouTube. In brief, the students' big motivation in learning pronunciation to some extent could not be considered as one of factors of the pronunciation errors.

Discussion
This study reveals that students of the English Education study program at Universitas Bengkulu still mispronounced some words containing inert letters. The students likely pronounced the inert letters that had to be silent and unpronounced the letters which had to be sounded. It is reflected by the most frequent errors which were predicated as very high error according to Sugiyono (2008). The errors occurred in the sounded inert letter /d/ of /-nd/ by 94% and 88% for silent inert letter /a/ of /-cal-/ The total percentage of the students' inert letter incorrect pronunciations is 36.7% and predicated as low error. It showed that the incorrect pronunciations were still fewer than the correct ones. This is in line with Ayumi (2018) that examined how the students of an English course pronounced the selected words with silent letter "g". The percentage of correct pronunciation from her research was 71.3% so the percentage of the incorrect ones was 28.7%. It showed that the percentage of the advanced level students' pronunciations that correctly silenced the silent letter was higher than the incorrect one. The results were similar because both of the research subjects were students who currently learned English intensively.
Contrarily, the result of this research has different result from Utama (2018). His study revealed that the pronunciation errors made by students were 162 of 240 total pronunciations. It showed that the incorrect pronunciations were higher than the correct ones. It is might due to the different subjects involved in both of the studies. This study involved English Education Study Program students meanwhile his study involved non-English Department students.
The highest percentages of the incorrect pronunciations were 94% for the sounded inert letter /d/ of /-nd/ and 88% for silent inert letter /a/ of /-cal-/. This is contrary with Alqunayeer (2016) which examined the pronunciation error of the letter /g/ in English written words by Saudi female learners who currently learned English. His research showed that 68% of students mispronounced /g/ before nasal which can be also categorized as inert letter. It was the highest percentage of all the pronunciations checked in his research. Meanwhile in this research, there were only averagely 35% of students who mispronounced /g/ before nasal /n/. It is probably happened due to the different focus of the research. This research focused on the inert letters pronunciations meanwhile his research focused on the /g/ pronunciations. Another reason why the result was not similar is this study was conducted online so the students had a chance to learn the words list given before making the video recording.
There were three factors of the inert letters pronunciation errors made by English Education students based on the interview data of this research. The first one was native language. All the students' native languages were not English. Their native languages were Bahasa Indonesia, Serawai, Lembak, Javanese, Pasma, and Rejang language. Only one of them had a rule like inert letters in English. This is in line with Utama (2018) and Ayumi (2018) that stated that the factor of the silent letter pronunciation errors was Indonesian phonology system which was different from English. They analyzed the rules of syllable structure and consonant cluster of both of the languages meanwhile the researcher of this research interviewed the students to find out whether the rule existed or not in their native language.
The second factor was age. Almost all the students started learning English at above 6 years old. It means that they missed their golden age to learn English. This is in line with Kenworthy (1987) which assumed that if someone does not have native-like pronunciation, they probably did not start learning English at their early age. It might happen since English education in Indonesia was implemented since the elementary school.
The last factor of students' pronunciation errors was amount of exposure. Most of the students claimed that they rarely found English in their daily social life. It means that they had lack of exposure to English. This result is in line with Kenworthy (1987) which stated that the learners who are surrounded by English and constantly exposed should affect their pronunciation skills. It could happen because the students live in EFL country which people in the country very seldom used English in their daily conversation.

Conclusion
Based on the result of the research, the researcher concludes that the most frequent errors of inert letters pronunciation by English Education Study Program at Universitas Bengkulu are sounded inert letter /d/ of /-nd/ pronunciation and silent inert letter /a/ of /-cal-/ pronunciation. The pronunciation errors are caused by three factors. The factors are native language, age factor, and amount of exposure to English.