Intelligibility and Comprehensibility of the Filipino English Accent to Hong Kong English Speakers

Intercultural communication between Hong Kong people and Filipinos is common and crucial in Hong Kong as Filipinos are the largest non-local ethnic group, and Filipino domestic helpers play a significant role in households and childcare. This study investigated the phonological patterns of Filipino English and explored the intelligibility and comprehensibility of the English spoken by Filipino domestic helpers to student teachers of English in Hong Kong. By analysing the speech produced by three Filipino English speakers, features of Filipino-accented English were identified. Moreover, by conducting listening-dictation tasks, it was found that the listeners considered 30% and 80.6% of the isolated words and keywords in a passage respectively to be intelligible, and 74.7% of the speech produced by Filipino English speakers was comprehensible. The major types of features of the Filipino English accent, which are also found to be factors which contribute to problems in intelligibility and comprehensibility such as consonant substitutions, deletion of consonants and consonant clusters, no distinction between long and short vowels, replacement of vowels, and shifting of word stress, are discussed in detail in this research paper along with the pedagogical and social implications.


INTRODUCTION
In the multicultural society of Hong Kong, Filipino is the largest non-local ethnic group, making up 2.5% of Hong Kong's total population in 2016 (Census & Statistics Department 2017). Most Filipinos in Hong Kong work as domestic helpers in households due to the large demand for caretakers in Hong Kong families because it is common for parents in Hong Kong to be so occupied by work that they have limited time to take care of their children. It is found that parents prefer domestic helpers over child care centres to take care of their children when family members are unavailable to do so (Tam 2001, Yeoh & Huang 1999. Thus, children are most likely to spend most of their time with domestic helpers. In fact, Filipino domestic helpers perform a significant role in households as they are often expected to be substitute mothers who nurture the children and are aware of what they need (Tam 1999). Therefore, there is a critical need for effective communication between domestic helpers and children, in which the children need to understand what the domestic helper says in order to establish a good mutual relationship and for the children and helpers to give feedback effectively to each other when expressing their needs. According to Tse et al.'s (2009) study, English is the medium of communication between 64% of 4,000 primary four children and their foreign domestic helpers in Hong Kong, indicating that English is crucial for mutual understanding and communication between most Hong Kong families and their domestic workers. In order to understand the effectiveness of communication between Hong Kong and Filipino English speakers, we explore the intelligibility and comprehensibility of understanding the other words provided in the same context. From another perspective, an individual may attain higher comprehensibility if a higher level of intelligibility is attained. The obvious positive relationship between intelligibility and comprehensibility shows the equal significance of investigating both intelligibility and comprehensibility. As mentioned in the introduction, high intelligibility of Filipino English at the word and sentence levels is important for Hong Kong speakers of English to obtain specific information and extract the meaning of sentences from recognising and understanding keywords in daily conversation with Filipino English speakers. Moreover, Chen (2011) found that Filipino listeners found it difficult to comprehend Hong Kong Cantonese speakers' speech well and vice versa. Accordingly, achievement of a high level of comprehensibility of Filipino English provides Hong Kong people with a basis for better communication and cooperation with the large population of Filipino English speakers in Hong Kong.

PHONOLOGICAL FEATURES OF FILIPINO ENGLISH
Previous studies such as those conducted by Tayao (2004), Dayag (2007), and Bautista and Bolton (2008) explicitly present the phonological features of Filipino English, including the following: substitution of the interdental fricative /ð/ by the alveolar stop /d/ and the interdental fricative /θ/ by the alveolar stop /t/; the voiceless labiodental fricative /f/ and the voiced labiodental fricative /v/ with /p/ and /b/; confusion of consonants such as /d/ and /t/; pronunciation of the aspirated plosives /p/, /t/, and /k/ as unaspirated plosives in a stressed syllable and word-initial positions; elimination of the ending sound in consonant clusters; shifting word stress to a different syllable; and the interchangeable use of vowels of contrasting lengths such as /i/ and /i:/ and /ʊ/ and /u:/. Although the phonological features of Filipino-accented English have been explored in past studies, language may change over time and variations of the phonological features in the existing database may be detected.

IMPACTS OF FILIPINO DOMESTIC WORKERS ON ESL LANGUAGE DEVELOPMENT
Multiple studies shed light on the positive influence of English-speaking Filipino domestic workers on children's English as a second language (ESL) language development. Improved performance in vocabulary identification has been found in Hong Kong Cantonese-speaking children who are cared for by an English-speaking Filipino helper than those who are not (Chan & McBride-Chang 2015, Dulay, Tong & McBride 2017. In addition, according to Tse et al. (2009), higher reading comprehension skills are demonstrated by primary school children who use English to interact with their foreign domestic helpers in settings where the parents are literate or illiterate in English. Moreover, people aged from 2.5 to 25 years old, including primary and secondary students who grew up with Filipino domestic helpers, have a greater ability to perceive and understand English in a Filipino accent than those who do not (Leung 2011, Leung 2012. Thus, previous research suggests that communication with Filipino English speakers may contribute to children's language learning, and if adequate support is provided to children in the classroom to learn about Filipino-accented English, the benefit may further extend to those who are not cared for by Filipino domestic helpers.

PAST RESEARCH ON ENGLISH AS AN INTERNATIONAL LANGUAGE
English often serves as a common language between two groups of L2 speakers with different L1s, including Filipino and Cantonese speakers in Hong Kong. From the "lingua franca core," introduced by Jenkins (2000), aspects of pronunciation that affect intelligibility can be deduced, including incorrect pronunciation of all consonants, failure in pronouncing contrasting vowel lengths, deleting consonants, and incorrect placement of word stress. Non-native speakers of English may benefit from Jenkins' (2000) pronunciation core by noticing the pronunciation features that affect intelligibility and further make adjustments in order to establish effective communication. However, the phonological features of Filipino English may differ from other varieties of English. At the same time, the phonological features that contribute to intelligibility problems for Cantonese speakers may not necessarily align with those for speakers of other native languages. For example, Abdely and Yap (2016) also found that Iraqi speakers of English encounter varying degrees of difficulty in English vowels /ɒ/ and /ae/. Sridhanyarant (2017) found that Thai speakers have difficulties in producing marked English fricatives /v/, /z/, /θ/, /ð/, and /ʒ/. Only the advanced learners could acquire unmarked /s/ and /f/ as well as marked /ʃ/ both initially and finally, and thus make their speech less intelligible. Thus, research focusing solely on the Hong Kong and Filipino English speakers may better represent the situation in Hong Kong.

COMPARISON OF INTELLIGIBITY AND COMPREHENSIBILITY BEWTEEN CHINESE AND FILIPINO ENGLISH SPEAKERS
The intelligibility of the English spoken by Chinese people to Filipinos has been explored in past studies, but the intelligibility of Filipino-accented English for Chinese English speakers is rarely mentioned. Chen's (2011) study shows that listeners from different backgrounds, including ten Filipinos, found 70% of the English spoken by Cantonese speakers in Hong Kong to be intelligible through a listening-and-dictation task of keywords in a passage. However, there is a lack of investigation on the intelligibility of Filipino-accented English for Hong Kong listeners. One of the few studies related to this area was conducted by Dayag (2007), whose research is on intelligibility. Spontaneous speech by five Filipino English speakers was recorded and played to six listeners from other countries for them to dictate back. The transcripts were later compared to the sample transcription made by the researcher. The results showed that listeners from Kachru's expanding circle, including one Mainland Chinese and one South Korean speaker, considered 55% of the Filipino English to be intelligible. However, the results of the study may not accurately reflect intelligibility because replacements of keywords with synonyms are marked as correct in that study. In fact, the use of synonyms only indicates the listener's ability to understand the meaning of the utterance instead of recognizing the word, and with reference to the definitions of intelligibility and comprehensibility mentioned previously in the literature review (e.g. Smith and Nelson (1985),) the results reflect comprehensibility instead of intelligibility. Dayag's study is also limited in terms of the variety of listeners as there were no Cantonese speakers from Hong Kong. Thus, more research is needed to understand both the intelligibility and comprehensibility of Filipino English to Hong Kong speakers of English. A common ground can be established for this study from the literature review. Firstly, intelligibility and comprehensibility are different but related concepts. Intelligibility refers to word perception, and comprehensibility refers to deriving meaning. Intelligibility at the word and sentence levels as well as comprehensibility are worth investigating because individuals have to both recognize and understand words in order to achieve effective communication. Secondly, exposure to and learning of the Filipino accent are beneficial as L2 language development may be fostered. Last but not least, the lack of research about the current phonological features of Filipino English as well as how these features affect the intelligibility and comprehensibility of Filipino English for Hong Kong English speakers indicates a research gap. Therefore, more information is needed to understand the level of effectiveness of communication between Hong Kong and Filipino English speakers so as to develop methods to enhance this communication.
The purpose of this research is to investigate the phonological patterns of Filipino English and explore the intelligibility and comprehensibility of the English spoken by Filipino domestic helpers to student teachers of English in Hong Kong. Student teachers were chosen as the target group because it is especially essential for future English teachers to understand features of Filipino English as they may need to communicate well with young students' domestic helpers about their academic performance and behaviour so that the message can be further relayed to the parents for better parent-teacher coordination. Additionally, English teachers serve as an important source of support for children's exposure to and understanding of Filipino English so that communication between children and Filipino domestic helpers at home can be further enhanced.
This research aims to answer the following research questions: 1. What are the phonological features of Filipino English speakers who are domestic workers in Hong Kong? 2. Do the phonological features of Filipino English affect the intelligibility and comprehensibility of utterances for student teachers of English in Hong Kong? If so, what are the phonological features that contribute most to problems in intelligibility and comprehensibility?
By analysing the speech produced by Filipino English speakers, the results of listening-dictation, the phonological features of Filipino English, and those features that contribute to the decreased intelligibility and comprehensibility to English speakers in Hong Kong can be identified. It is predicted that the phonological features of Filipino English reflected in previous research may appear along with slight variations. The dictation results may reflect several phonological features that contribute to failures in intelligibility and comprehensibility, which will serve as a reference in the development of adaptation strategies.

PARTICIPANTS
Three speakers and 27 listeners were involved in this study. The three speakers, aged 41 to 48, were born and raised in the Philippines (detailed information in Appendix D). Three of the speakers fulfilled the selection criteria in that they were (1) Filipino, (2) able to read and speak English fluently, (3) working as domestic workers in Hong Kong, and (4) currently working or had worked as caretakers of children. The utterances produced by the three speakers for the items in the word list, passage, and speech were screened, and the utterances that best represented the features of the Filipino accent, such as substitutions of /ð/ by /d/and /θ/ by /t/, were selected and compiled together to construct the audio for the listeningdictation tasks.
The 27 listeners, aged 19 to 22, were student teachers majoring in Bachelors of Education in the English Language. The selection criteria for the listeners were (1) currently obtaining an education degree in the English language, (2) Hong Kong residents who had lived in Hong Kong for more than seven years and (3) Level 4 in the Hong Kong Diploma of Secondary Education (HKDSE) or 6 in the International English Language Testing System. Table 1 summarises the selection criteria of the participants in this study. By setting the above controlling criteria, the results can focus on describing the intelligibility of Filipinoaccented English to future English teachers in Hong Kong by avoiding large variations due to differences in mother tongue influence and language proficiency. The tools involved in the data collecting process included a list of words and a reading passage for the speakers to read aloud in order to understand the intelligibility of Filipino English at the word and sentence levels to the student teachers of English. Additionally, a topic of speech that was accompanied by a few images for a specific theme was given to the speakers of Filipino English for them to produce a speech 1.5-5 minutes long so as to understand the comprehensibility of Filipino English in a relatively natural speech that is closer to communication in daily life.
In the first task (Appendix A1), the three speakers were asked to pronounce a list of 105 words in order to investigate the intelligibility of Filipino English at the word level. The main advantage of the word list is its focus on the special phonological features of Filipino English that may create difficulty in understanding, for example, words with fricatives, including /ð/ (e.g., weather), /θ/ (e.g., thorn), /f/ (e.g., far), and /v/ (e.g., vest), words in which /p/, /t/, or /k/ is in the stressed syllable (e.g., suppose, potato, attack), words ending with consonant clusters (e.g., west), words with two or more syllables (e.g., broccoli), and words including vowels of contrasting length (e.g., seek and sick). Listeners were then invited to write the words as dictation on the answer sheet (Appendix A2). Word intelligibility was measured by the correct number of words over the total number of words in the test. Phonological analysis of the audio recording was done in terms of vowels, consonants, and word stress. After that, the answers from the listeners were marked. Any answers that were spelled incorrectly in task 1, including words with other meanings and those whose spelling was not similar to any word in the dictionary were counted as wrong because they showed that the listener was unable to identify the word. However, the spelling of homophones, such as spelling "very" as "vary" and "whether" as "weather" as well as words that are pronounced identically in British-accented English and American-accented English were counted as correct because it was difficult to determine which word was used when the words were isolated from a context, and listeners were not informed whether the audio obeyed the rules of the American or British accent.
In the second task, a reading passage comprising 12 sentences was given to the three speakers for them to read aloud (Appendix B1). After that, listeners were invited to fill in 33 blanks in these sentences (Appendix B2) so that the intelligibility of Filipino English at the sentence level could be investigated through the number of keywords correctly identified by the listeners. The advantages of this test are that firstly, the use of a set script allows the inclusion of a fixed set of vocabulary that represents the special phonological features of Filipino English. Secondly, the sentences were deliberately chosen and refined in order to prevent repetitive sentences and the incorrect use of grammar and fillers that were most likely redundant in expressing meaning and difficult for listeners to syllabify and take as dictation. Thirdly, the use of a gap-fill instead of full sentence dictation enables listeners to focus on memorizing and spelling the content words, which are the major factors of intelligibility at the sentence level. Major mistakes in spelling from the listeners' dictation, for example, writing "toxic" as "toxin" were counted as wrong as the former is an adjective while the latter is a noun, and their pronunciation is significantly different.
The third task (Appendix C1) involved asking the three speakers to each produce a 1.5-5 minute speech. The listeners then referred to the selected speech to answer true/false comprehension questions (Appendix C2). First, the use of speech in assessing comprehensibility allows a more authentic representation of spontaneous speech in daily communication. Second, a given topic of the speech is guided along with certain images and a storyline so that the utterances will be more logical and to prevent factors of failure in comprehension other than pronunciation. Third, keywords were specifically chosen so that the words that better express the phonological features of Filipino English could not be intentionally avoided by the speakers. Fourth, the use of true/false questions allows listeners to show their understanding of the meaning of utterances through the correct judgement of answers. Therefore, the comprehensibility of Filipino English to student teachers of English in Hong Kong may be investigated through the speech comprehension task. In task 3, wrong selections and blank answers were counted as wrong as they do not prove that the listeners were able to comprehend the speech and determine whether the statement was true or not. The intelligibility and comprehensibility scores were calculated by the number of correct answers over the total number of items. The dictation results were then analysed and compared to the errors or features produced by the speakers in the audio in order to understand whether the features affected the intelligibility and comprehensibility to the student teachers. Figure 1 shown below illustrates the sequence of research procedures. The process and aim of the study were explained to the listeners before they performed the listening and dictation tasks. Three Filipino English speakers were asked to perform the three speaking tasks. Each of them read the word list and then the passage, followed by the presentation of a speech. Time was given for the speakers to prepare the tasks. The speeches were recorded with an Apple iPhone 6s (Cupertino, CA, USA) located about 3.5 inches away from the speakers' lips using the Voice Memos application at 16-bit pulse-code modulation and a sampling rate of 44 KHz. The audio files and answer sheets were sent by email and WhatsApp to the listeners who were invited and had agreed to participate. The instructions were then explained to listeners; for example, they could listen to the audio one time only, and they had to stop the audio when they heard a "beep" sound so that they could type their answers on the answer sheet. The participants then sent back the completed answer sheets. A consent form was signed by each of the speakers and listeners involved in the study. The texts of the dictation and answers for the comprehension questions were then analysed together with the recording of the speeches.

RESULTS
In this section, outcomes from the phonological analysis of the recorded speech will first be presented and discussed, followed by the results from the word list dictation, dictation of keywords in sentences, and comprehension task. Error rates and error patterns will then be presented.

PHONOLOGICAL ANALYSIS OF FILIPINO-ACCENTED SPEECH
The phonological analysis of the recordings of the three Filipino English speakers was conducted in terms of consonants, vowels, and word stress. Table 2 shows the phonological features of the Filipino English accent for consonants, including word-initial, word-medial, and word-final consonant substitution, consonant or consonant cluster deletion, and final consonant substitution after deletion. (2) Dictation of words in passage (3) Completion of comprehension task according to speech of words in passage Analysis of how pronunciation in the recordings may affect the score of three tasks respectively The three recordings and answer sheets were sent to student teachers of English through email and Whatsapp Completed dictation and comprehension answer sheets were sent back from the listeners  Table 3 shows the special features of vowel pronunciation in Filipino-accented English such as confusion of long and short vowels, vowel shortening, replacement of schwas and diphthongs by short vowels, replacement of short vowels by long vowels, and replacement of short vowels by other short vowels.  Table 4 shows the special features of word stress in Filipino English, including moving the word stress from the front to the second syllable in three-syllable words, moving the front stress to the second or third syllable in words with four syllables, and switching the stress from the final to the front syllable in two-syllable words. The answers from the 27 listeners, who are student teachers in Hong Kong, were marked and their scores were calculated by considering one correct item/word as one mark; the full mark for the word list diction (task 1) was 105. The average score is measured by addition of the total marks of the 27 answer papers, which was 827, followed by dividing the total mark by 27, which is 31.4 (standard deviation=6.2). The intelligibility of the Filipino accent at the word level was then derived by dividing the average score (31.4) by the total score (105) and multiply by 100%, leading to the finding of intelligibility at about 30%. This shows that the listeners were not able to identify the majority of the words in the Filipino accent, implying that they may have difficulty in understanding these words if they are included in daily conversation, which may affect the effectiveness of communication with Filipino domestic workers.
The error rate of each word in the word list is calculated through dividing the total number of listeners who failed to dictate the correct word in the dictation task by the total number of listeners, and multiplied by one hundred percent. Table 5 explains the most frequent words that the listeners failed to recognize and write down in the word list dictation. The 73 words included in the table were unrecognized by more than half of the listeners. There are several possible issues contributing to the unintelligibility of the Filipino-accented English words to the listeners. First of all, the accents of the three Filipino domestic helpers may be too strong for the listeners to decode the words correctly, and it is difficult to guess the words in an isolated context. Secondly, the listening skills and vocabulary knowledge of the listeners may also be a factor in writing a wrong answer, especially for the words that are less commonly used in informal speech, for example the words "grit," "advocacy," and "cemetery," which the student teachers were possibly unfamiliar with. Frequent dictation errors were analysed with the features of the Filipino English accent found in the audio recordings of the word list to understand the distinctive features of Filipino-accented English that may contribute to the failure in intelligibility. Table 6 presents the frequent types of features that caused errors in dictation, including consonant substitution at the initial, medial, and final positions in words, no distinction of long and short vowels, vowel shortening, vowel substitution, replacement of word stress from the first to the second or third syllable in four-syllable words, as well as the switching of word stress from the second syllable to the first in two-syllable words.

RESULTS OF THE DICTATION OF KEYWORDS IN A PASSAGE (TASK 2)
In the second task, listeners were asked to fill in the blanks of a short passage, where the blanks were the keywords that conveyed important meaning in the sentences. There were a total of 33 blanks, and each correct answer was counted as one mark. The average score was 26.6/33, with a correct answer rate of 80.6% (standard deviation=3.3). The intelligibility rate is much higher than that of the word list dictation. Several possible reasons may contribute to this difference. Firstly, the variation of keywords in the passage was not as large as the words in the wordlist, so it was easier for listeners to perform well in task 2. Secondly, the provision of text in other parts of the passage (except the keywords) may have allowed the listeners to understand the context and predict the words correctly despite any confusion that may have been caused by unfamiliar pronunciation features. The error rate of each keyword in the fill-in-the-blanks dictation is calculated through dividing the total number of listeners who provided the wrong answer in the dictation task by the overall number of listeners, and multiplied by one hundred percent. Table 7 displays the most frequent errors made by more than half of the listeners in task 2. These errors may be related the listeners' unfamiliarity with the strong Filipino accent or a limitation in vocabulary that caused the listeners to be unable to write the correct keyword, even in a known context of provided text.  Table 8 displays the types of pronunciation errors that contributed to the unintelligibility of the keywords to the listeners. The types of errors that were also found to be frequent causes of unintelligibility in task 1 included consonant shifts from voiceless to voiced plosives, as in /k/ to /g/, voiceless fricative /f/ and voiceless plosive /p/ to voiced plosive /b/, and substitution of short vowels by another short vowel, for example /e/ and /əә/ to /ɪ/. In addition, other examples of pronunciation errors that could be possible causes of unintelligibility were found, such as replacing short vowels by long vowels, for example, /ʌ/ by /ɑː/ and /e/ by /ɝː/. TABLE 8. Frequent error types/features in fill-in-the-blanks in passage dictation mentioned in the speech (transcript in Appendix E). The error rate of each question is calculated by number of listeners who provided the wrong answer divided by the total number of listeners. Table 9 shows the most frequent three questions which were answered incorrectly by the listeners. More than half of the listeners were unable to comprehend that the main characters in the speech-the little duck and the mother duck-were real ducks instead of rubber toys. About 44% of the listeners could not comprehend the incident that happened as a plot twist in the story-the duck's mother used tape to fix the human doll's head, and 37% of the listeners were unable to understand that the story involved a human doll and that only the mother blew air into the human doll while the little duck just sat in the bathtub crying, which was also part of the climax in the story. The listeners' comprehensibility rating of the speech may be affected by multiple factors. First of all, the listeners may be unfamiliar with the Filipino-English accent. Also, minor grammatical mistakes such as inconsistencies of tense and pronouns in referring to the little duck may have caused confusion and affected the listeners' understanding. In addition, the inadequate listening skills of the listeners, for example, being unable to pay attention to key points, and a lack of reading skills that made them unable to understand the comprehension questions may also be causes of incorrect answers. As this study focuses on how Filipino English pronunciation may affect comprehensibility for student English teachers in Hong Kong, the frequent pronunciation errors of keywords in the speech are analysed and presented in Table 10. The frequent types of pronunciation errors that were also found in tasks 1 and 2 include consonant shifts from /k/ to /g/ and /p/ to /b/. Other than that, deletion of the consonant /h/ and replacement of the vowels /e/ by /ae/ and /əә/ by /ʌ/ were also causes of comprehensibility issues. Regarding consonant substitution, the voiceless plosives /p/, /t/, and /k/ were voiced in the word-initial position, which is in-line with Dayag's (2007) research. However, the difference in the findings of this study is that the pattern did not occur only in the word-initial position but also in the medial and final positions of words, even when they were not in the stressed syllable. Additionally, the voiceless plosive /p/ was substituted by the voiceless fricative /f/ at the word-initial position, suggesting a new pattern of voiceless stop substitution in Filipino-accented English. In addition, the substitution of fricatives by plosives at word-medial positions aligned with the results of previous research, where the voiced fricative was often substituted by the voiced plosive, as in /ð/ to /d/ and /v/ to /b/, while the voiceless fricative was often substituted by the voiceless plosive, as in /θ/ to /t/ and /f/ to /p/ (Dayag 2007, Tayao, 2004, but the pattern was slightly different in word-initial and word-final positions, where voiced and voiceless fricatives were substituted by voiced plosives in some words and voiceless plosives in others. Moreover, a new pattern of voiced affricatives substituted by voiceless affricatives at the initial and final position of words was found in this study. In addition to consonant substitution, there were deletions of the last consonants of consonant clusters at the final position of words, similar to Tayao's (2004) findings. However, some consonant clusters were eliminated and replaced by other consonants as well, presenting a feature of Filipino English that is absent from past studies.
In terms of vowel pronunciation, there was no distinction between long and short vowels, matching Bautista and Bolton's (2008) description. The results of the study also displayed new patterns such as vowel substitution, including the replacement of schwas and diphthongs by short vowels, short vowels by long vowels, and long vowels by diphthongs.
Word stress was also found to be relocated from the first to the second syllable in words that consist of three syllables, which is similar to the results in Tayao's (2004) study. Despite this, the results of the study also demonstrated the relocation of word stress from the first to the second or third syllable in four-syllable words, and the switching of the front and ending stress in disyllabic words.
The second research question, "Do the phonological features of Filipino English affect the intelligibility and comprehensibility of utterances to student teachers of English in Hong Kong?" and the follow up question "If so, what are the critical phonological features that contribute to problems in intelligibility and comprehensibility?" were answered through the results from the dictation and comprehension tasks. It was found that that 30% of the isolated words were intelligible to the listeners, 80.6% of words in a passage were intelligible, and 74.7% of the speech was comprehensible. The phonological features that were mentioned in the earlier part of the discussion, including consonant substitution, consonant or consonant cluster deletion, no distinction of long and short vowels, replacement of vowels, and displacement of word stress were all found to be critical variables that were related to errors in dictation and contributed to problems in intelligibility while the substitution of consonants and vowels and deletion of consonants were the major pronunciation features related to problems in achieving comprehensibility.
There are a few limitations to this study that can be adjusted in future follow-up studies. Firstly, although the rationale behind the word list dictation is to understand the intelligibility of isolated words spoken in the Filipino English accent to Hong Kong English speakers, it seemed that requiring listeners to dictate isolated words without the support of a context may have limited listeners' opportunity to make use of possible adaptation strategies to recognize the words accurately. The task could be adjusted to include questions on the answer paper, and listeners would have to listen to the words produced by Filipino English speakers, where the words are answers related to the respective questions on a variety of topics, for example, "What is the mentioned place?" In this way, the range of possible answers would be narrowed to words relevant to the topic so listeners can get an idea about what kinds of words they are expected to write, and they can make use of adaptation strategies to adjust the features of Filipino English in their minds in order to identify the words accurately. Secondly, due to time limitations, the listeners were allowed to complete the listening tasks by themselves at home. This could limit the control of variables such as background noise, which may distract the listeners and affect their accuracy in hearing and completing the dictation and comprehension tasks accurately. Therefore, future studies should invite the listeners to complete the tasks in a quiet venue so that distractions may be avoided for more reliable results.

CONCLUSION
In conclusion, the phonological features of the Filipino English accent, the comprehensibility and intelligibility of Filipino English to Hong Kong English speakers, and the features of Filipino English that affect intelligibility and comprehensibility were investigated. The phonological patterns of Filipino-accented English detected in this study included consonant shifts, deletion of consonants and consonant clusters, no distinction between long and short vowels, replacement of vowels, shifting of the stressed syllable from the first to the second syllable in three-syllable words and from the first to the second or third syllable in foursyllable words, and switching the front and ending stress in two-syllable words. In addition, it was found that Hong Kong speakers of English who are student English teachers found 30% of the isolated words and 80.6% of words in a passage to be intelligible, 74.7% of speech produced by Filipino English speakers was found to be comprehensible, and all of the phonological features of Filipino English discussed were found to cause a certain amount of failure in intelligibility whereas the substitution of consonant and vowels and deletion of consonants are the major pronunciation features related to failures in comprehensibility. Therefore, Hong Kong English speakers should understand more about the phonological features of Filipino English, which will lead to a high level of intelligibility and comprehensibility of the utterances spoken by Filipino English speakers, whom they may often encounter in daily life, for better understanding and cooperation.
There are several implications of the study in social, cultural, and pedagogical aspects. Hong Kong speakers of English may use as a reference the special phonological features of Filipino-accented English discussed in the study and develop adaptation strategies in order to decode keywords in sentences more accurately in daily conversation with the large population of Filipinos in Hong Kong, especially in homes where Filipino domestic helpers are working. The findings of this study may also help Hong Kong speakers to obtain information from isolated words more accurately on the phone for emergency, bank, and telecommunication services as well as recruitment procedures.
In terms of cultural exchange in intercultural communication, a better understanding of the features of Filipino-accented English through the results of this study may allow the utterances of Filipino English speakers to achieve higher intelligibility and comprehensibility to Hong Kong speakers of English so that they may have deeper conversations with lowered language barriers and better understand the culture and traditions of Filipino people.
From an educational perspective, teachers can use as a reference the phonological features of Filipino-accented English discussed in the study that may cause difficulty in intelligibility and comprehension in order to educate their students about the Filipino variety of English; studies have proven that more exposure to Filipino-accented English may allow children to better understand Filipino English (Leung 2011, Leung 2012. This is beneficial when children communicate with Filipino domestic helpers at home or when they encounter Filipino speakers of English in exchange programs abroad and even in business meetings in their future careers. Teachers may also make use of the features of Filipino English discovered in this study to teach the Filipino variety of English as several studies have shown that greater exposure to Filipino-accented English may benefit children's language development in terms of vocabulary identification (Chan & McBride-Chang 2015, Dulay et al. 2017) and reading comprehension skills (Tse et al. 2009). Moreover, the result that the intelligibility of isolated words is significantly lower compared to words in context dictation indicates that higher intelligibility can be achieved when textual and contextual aids are provided even though listeners are unable to recognize isolated words in Filipino-accented speech. Thus, English teachers may introduce the Filipino variety of English to their students by focusing on specific contexts and topics, for example, environmental issues, along with the support of audio and print that would enable students to compare unfamiliar pronunciation to other texts. When students are familiarized with Filipino-accented English and are able to make use of adaptation strategies to perceive higher intelligibility and comprehensibility, they may communicate better with Filipino English speakers.
In addition, the results of the study may help improve childcare services provided by Filipino domestic workers. Many students at the primary school level are cared for by domestic helpers from the Philippines, and teachers may need to discuss important information such as student behaviour and performance at school with Filipino domestic workers so that they can relay this information to the students' parents. In this case, teachers may also take notice of the special phonological features of Filipino English discussed in the study to adapt to the foreign pronunciation accordingly in their minds so as to understand the utterances of the Filipino domestic helpers in their conversation. The researched outcomes may also help children to understand the utterances of their Filipino helpers so that accidents and disputes in their daily lives due to miscommunication may be prevented and may help the children to build a closer relationship with their Filipino domestic helpers, who often substitute in the mothers' role.
Last but not least, this study may offer information and insight for future research that aims for a deeper understanding of the Filipino English accent and the development of a common core of phonological features between Cantonese-accented and Filipino-accented English so that Filipino English speakers' intelligibility and comprehensibility to Hong Kong speakers may be improved and further foster economic development, social security, and social harmony as well as the quality of childcare services provided to children in Hong Kong.