Intercultural Pronunciation Variations and Adjustment Strategies among EFL Students from Java, Nusa Tenggara, and Papua

We propose a new approach for performing phonetic transcription of text that utilizes automatic speech recognition (ASR) to help traditional grapheme-to-phoneme (G2P) techniques. This approach was applied to transcribe Chinese text into Taiwanese phonetic symbols. By augmenting the text with speech and using automatic speech recognition with a sausage searching net constructed from multiple pronunciations of text, we are able to reduce the error rate of phonetic transcription. Using a pronunciation lexicon with multiple pronunciations for each item, a transcription error rate of 12.74% was achieved. Further improvement can be achieved by adapting the pronunciation lexicon with pronunciation variation (PV) rules derived manually from corrected transcription in a speech corpus. The PV rules can be categorized into two kinds: knowledge-based and data-driven rules. By incorporating the PV rules, an error rate of 10.56% could be achieved. Although this technique was developed for Taiwanese speech, it could easily be adapted to other Chinese spoken languages or dialects.

A method of accent modeling through Pronunciation Dictionary Adaptation (PDA) is presented. We derive the pronunciation variation between canonical speaker groups and accent groups and add an encoding of the differences to a canonical dictionary to create a new, adapted dictionary that reflects the accent characteristics. The pronunciation variation information is then integrated with acoustic and language models into a one-pass search framework. It is assumed that acoustic deviation and pronunciation variation are independent but complementary phenomena that cause poor performance among accented speakers. Therefore, MLLR, an efficient model adaptation technique, is also presented both alone and in combination with PDA. It is shown that when PDA, MLLR and PDA+MLLR are used, error rate reductions of 13.9%, 24.1% and 28.4% respectively are achieved.

In foreign language learning, understanding is not limited to basic sounds such as vowels and consonants, but also includes suprasegmental features that influence meaning and clarity in communication. One important suprasegmental feature is word stress. For Indonesian learners of German, applying correct word stress often presents a particular challenge. This study aims to examine how learners perceive and produce word stress in the German language. A descriptive qualitative methode with an auditive phonetic approach was employed to investigate the perception and production of word stress as a suprasegmental element in German. Data were collected in two stages: (1) a production task involving the oral reading of words and pseudowords, which were analyzed auditorily, and (2) a perception task in which participants listened to recorded words and identified the position of stress. The analysis compared the accuracy of participants’ production and perception and identified common patterns of error. The findings are expected to serve as a basis for developing more contextual and learner-oriented suprasegmental phonetics teaching materials for Indonesian learners of German.

Use of natural languages for the computer communication is one of the current research topics. The speech recognition plays a central role in communicating the computer by means of speech. Speech Recognition is the process of converting analog signal into the symbolic gesture form known as text. An Automatic Speech Recognition (ASR) is the process of converting input speech signal given to the system, into text. This input signal is any human spoken word. Though the last Four decades, research is going on bringing the system perception near to the human being, in recognition word accuracy. Many domains play a role in degrading the of system performance in which language and its pronunciation variants caused by different reasons like accent, gender, dialects of the speech are in general factors. In specific, system environment, articulatory phonetics, acoustic phonetics, (acoustic system), lexical model (pronunciation dictionary) and language model including the mood of the speaker. Acoustic modeling can be done in most of the cases by using signal processing, where as lexicon model require a sufficient human intervention as, it is based on the language and human perception. Once sufficient primary data is built then automatic processing can be done using different modeling techniques to derive more data for proper training of the speech recognition system. Here the linguistics are play a major role in building the robust system. In lexical model design language plays a major role. Each language has its own rhythm in speech and language aspects. Based on language rhythm worldly languages are classified as stress timed and syllable timed rhythm. Most of ASR systems use the lexical model that is built for stressed timed languages. All Indian languages are syllable timed rhythmic languages one such is Telugu Language. In this paper analysis of the decoding results of ASR system using two different lexical model environments. One is CMU lexicon which is based on stress timed language as the tool is used American accent English phonemes and another UOH lexicon which is handcraft lexicon for Telugu language which is also a syllable timed language.. Further studied are the gender and accents (pronunciation variant factors) effecting the Substitutional errors in ASR system. The confusion matrix for vowel and consonants alone analyzed for both cases and also for isolated word recognition where the confusion matrix gives the most common phonemes substituted. In all the cases the UOH lexicon based ASR system gives the improvement of word accuracy around 20 to 30%. Speech is a process used to communicate from a speaker to listener. Pronunciation relates to speech, and humans have an intuitive feel for pronunciation. For instance, people chuckle when words are mispronounced and notice when foreign accent colors a speaker's pronunciations.(1). If the words were always pronounced in the same way, ASR would be relatively easy. However, for various reasons words are almost always pronounced differently and varied from one speaker to another and from once situation to another. The variability is due to co-articulation, reasonal accents, speaking rate, speaking style etc.

Compared to more familiar varieties of Swedish, the dialects spoken in Finland have rather diverse syllable structures. The distribution of distinctive syllable weight is determined by grammatical factors, and by varying effects of final consonant weightlessness. In turn it constrains several gemination processes which create derived superheavy syllables, in an unexpected way which provides evidence for an anti-neutralization constraint. Stratal OT, which integrates OT with Lexical Phonology, sheds light on these complex quantity systems. 1 The weight of stressed syllables 1.1 Light stressed syllables1 The bimoraic minimum: Sweden vs. Finland. In most Swedish dialects of Sweden (here referred to as West Swedish for short), stressed syllables are minimally bimoraic: they must contain at least a long vowel (-VV-) or a closed syllable (-VC-).2 Words like (1a) are therefore impossible. Because word-final consonants are weightless (“extrametrical”) in Swedish, the two-mora minimum also excludes monosyllabic words with -VC rhymes (see (1b)): (1) a. *[ro], *[ro.da], *[ro.a], *[no.gra] b. *[ro(d)] The Swedish dialects of Finland present a more varied picture. Only parts of Aland have the two-mora minimum (e.g. Brando and Kumlinge in the northeastern part of the island, Sundberg 1993:131 ff.). All other Fenno-Swedish dialects allow light (i.e. monomoraic) stressed syllables as a distinctive syllable type: (2) [daga], [dagar] ‘days’, [viku] ‘week’, [veliN] ‘gruel’, [suvel] ‘food eaten with bread, sowl’, [somar] ‘summer’, [stygu] ‘hut’, [paron] ‘potato(es)’, [hakon] ‘the chin’, [hole] ‘the hole’, [segla] ‘to sail’, [tala] ‘to talk’, [sita] ‘to sit’, [myky]3 ‘much’, [stad0gari] ‘steadier’, [snidit] ‘askew’, [Uyvu] ‘twenty’ Fenno-Swedish, then, has a lexical contrast between stressed CV, CVC, and CVV syllables:4 (3) a. [baka] ‘bake’ (99), [baaket] ‘after’ (adv.) (114), [bakkan] ‘the hill’ (114) b. [vaten] ‘water’ (102), [maaten] ‘the food’, [natten] ‘the night’ (70) c. [betar] ‘better’ (51), [fleetor] ‘braids’ (43), [tvettar] ‘washes’ (51) Even though stressed CV syllables are allowed, words of the form CV are categorically excluded in all the dialects (except for function words, on which see below).5 As for words of the form CVC, the dialects are divided. Most allow them:6 1The information on Fenno-Swedish dialects given here is based primarily on the 29 transcribed dialect texts in Harling-Kranck 1998, with accompanying tapes, as well as on the brief grammatical sketches of the dialects provided there. Page references below are to that work, unless otherwise specified. For supplementary information on particular points I have consulted the additional dialect monographs cited below. Special thanks are due to Mikael Reuter, for valuable discussion of Helsinki Swedish, and for generously providing me with a copy of his unpublished thesis (Reuter 1982). 2Except where otherwise stated, the generalizations stated here hold for phonological words. Each member of a compound constitutes a separate phonological word. 3Here and throughout I ignore dialectal variation in pronunciation where it is not relevant to the analysis of syllable weight. For instance, dialects with palatalization before front vowels have [myUy] or [myUi] instead of [myky]. 4In phonetic transcriptions of Fenno-Swedish, I adhere to IPA standards except that I mark vowel and consonant length by gemination, so as to conform with the phonological (lexical) representations, and to allow convenient marking of syllable boundaries (by “.”). Italics are reserved for citing word in Swedish spelling, which will be done for standard West Swedish and standard Helsinki Swedish only. 5The single contrary example is ga [ga] ‘go’ in Vora (central Ostrobothnia, Harling-Kranck 1998:121), apparently a fast speech variant of that dialect’s normal [gaa]. 6The contrast between /CVC/ and /CVCC/ is clearest before a vowel in close contact, e.g. [hol i mitten] ‘hole in the middle’, [r0nn o] ‘round too’ (H.-K. 22). The /CVC/ words are partly retentions of Proto-Nordic /CVC/, partly analogical reintroductions (Hulden 1957:122), and partly apocopated from CVCV at different periods.

This paper reports on a qualitative inquiry that investigated a group of Hong Kong university students’ perceptions of their linguistic identities in English as a Lingua Franca (ELF) communication. The participants of the study were proficient second language (L2) speakers of English who reported to have participated in intercultural ELF communication on a regular basis. By analyzing data gathered via two rounds of in-depth interviews, the study revealed that these participants did not evaluate their “non-native” status negatively but embraced their identities as legitimate and empowered speakers of English in ELF interactions. In addition, they validated their identities as multilingual, multicompetent, and translingual speakers of English with the ability to draw upon a diverse set of multilingual resources flexibly in ELF contexts. The analysis also found that these participants’ linguistic identities were constructed and negotiated in relation to the perceived linguistic competence of other ELF speakers, especially L2 speakers of other cultural/national backgrounds. This paper sheds light on the complexity of linguistic identities in ELF contexts and contributes to the growing ELF research in the Asian context.

The study assessed the clinical utility of an invented spelling tool and determined whether invented spelling with linguistic manipulation at segmental and supra-segmental levels can be used to better identify reading difficulties. We conducted linguistic manipulation by using real and nonreal words, incorporating word stress, alternating the order of consonants and vowels, and alternating the number of syllables. We recruited 60 third-grade students, of which half were typical readers and half were poor readers. The invented spelling task consistently differentiated those with reading difficulties from typical readers. It explained unique variance in conventional spelling, but not in word reading. Word stress explained unique variance in both word reading and conventional spelling, highlighting the importance of addressing phonological awareness at the supra-segmental level. Poor readers had poorer performance when spelling both real and nonreal words and demonstrated substantial difficulty in detecting word stress. Poor readers struggled with spelling words with double consonants at the beginning and ending of words, and performed worse on spelling two- and three-syllable words than typical readers. Practical implications for early identification and instruction are discussed.

We propose a method of incorporating pronunciation modeling into acoustic models with high discriminative power and low complexity to improve spontaneous speech recognition accuracy. Spontaneous speech contains a higher level of phonetic and acoustic confusions due to the larger degree of pronunciation variations caused by speaking rate, speaker style, speaking mode, speaker accent, etc. In general data-driven complexity-reduction methods without explicit modeling of pronunciation variations, the acoustic model is not robust enough to capture the flexible phonetic confusions and pronunciation variants in spontaneous speech. We propose a state-dependent phonetic tied-mixture (PTM) model with variable codebook size to improve the coverage of phonetic variations while maintaining model discriminative ability. Our state-dependent PTM model incorporates a state-level pronunciation model for better discrimination of phonetic and acoustic confusions, while reducing model complexity. Experimental results on the spontaneous speech part of Mandarin Broadcast News shows that our model outperforms state tying and mixture tying models by 2.46% and 3.51% absolute syllable error rate reduction, respectively, with comparable model complexity. After adding Gaussian sharing to the latter models, our proposed model still yields an additional 1% and 2.6% absolute syllable error rate reduction. In addition, unlike many complexity reduction methods, our method does not lead to any performance degradation on read speech.

Articulatory rate and pause time in a standardized reading task in Parkinson's disease (PD) patients in correlation to disease duration and severity as compared to healthy controls were analyzed. In 121 PD patients and 70 healthy controls, an acoustical analysis was performed on the first and last sentence of a standardized 170-syllabic text, using a commercial audio software. Articulatory rate and speech to pause ratios were calculated by measuring the length of each syllable and each pause both at the end of words and within polysyllabic words. No significant difference in overall articulatory rate was found between PD patients and controls. Both groups showed an accelerated speech rate in the last sentence compared to the first; however, PD patients had a higher speech acceleration than did controls. PD patients exhibited a significantly reduced percental pause duration in relation to total speech time in the first sentence and a reduced percental pause time within polysyllabic words. PD patients made significantly less but longer pauses at the end of words and less pauses within polysyllabic words. UPDRS III showed an inverse relation to number and rate of intraword pauses, and disease duration was negatively correlated with articulatory rate. The characteristics of parkinsonian speech feature was not only a stronger acceleration of articulation rate in the course of speaking but also a significant reduction of the total numbers of pauses, indicating an impaired speech rhythm and timing organization.

The dialect spoken in the small southern Appalachian community in Western North Carolina, formally a part of the Inland South, represents the most archaic features of Southern English, including monophthongization of the diphthong /ai/ to [a:]. While Southerners produce the monophthong before voiced consonants (“prize”) all across the South, pre-voiceless monophthongization (“price”) is more restricted and still occurs in the Inland South. However, as the local culture is gradually shifting toward new mainstream sociocultural norms, so does the local dialect. To document the sound change in progress, this acoustic study examined the /ai/-monophthongization across several generations of local speakers ranging from 8 to 91 year olds, 58 males and 60 females, using a style shifting paradigm. Older generations produced the monophthong irrespective of speaking style and consonant context. Young adults introduced the diphthongal variant in pre-voiceless context. The actual departure from the monophthong occurre...

The occurrence of phonological interference can be found if the native is able to acquire multilingually. The research aims to reveal the phonological aspects causing phonological interference and establish the strategies for avoiding interference. Students of the English literature department that take English Language Teaching (ELT) and speak Javanese, Indonesia, and English language become the primary sources of this research. The method used in this research is a descriptive qualitative as research design and the theory of Geoffry S. Nathan for the analysis. Meanwhile, Marsono’s theory is employed to compare it with vocal, diphthong, and consonant in Javanese language. The finding shows that phonological interference comes from a segmental aspect, totaling errors up to 99 and 189 in the suprasegmental aspect. In the segmental aspect, errors of English pronunciation were found on the assimilation of speech implementation based on vowel, diphthong, and consonant. Meanwhile, in the suprasegmental aspect, word stress on the second syllable with two, three, and four-syllable words become the most errors. Therefore, phonological interference can be avoided by doing more individual practice in looking for difficult vocabularies and using English as the main language in doing enjoyable activities such as listening to music and watching movies.

The main aim of this study is to find out the correlation between EFL students’ perceived linguistic proficiency and their level of confidence in using English. To achieve that, sixty first year students at the faculty of English language and literature, Ajloun University in Jordan, were selected based on convenience sampling. The subjects were asked to self-assess their linguistic proficiency using five-point Likert scale and to respond to the survey questionnaire. To test the statistical associations between the two variables, ANOVA was used to confirm if there were any statistically significant differences in self- assessment scores among the groups of students having different levels of confidence. To confirm where the differences lie, Post Hoc tests (Turkey HSD and Homogeneous Subsets) were used. NOVA gives an F value of 44.131 with 2 and 57 degrees of freedom (p < 0.01), showing that there are differences in self-assessment means among groups. The multiple comparisons among mean differences by the Post Hoc tests (Turkey HSD and Fisher’s LSD) show that the group of students who are 'hardly confident' are significantly different from the other two groups of students. The group of students who are 'a little confident' are significantly different from 'hardly confident' students and 'confident' students. Finally, the group of students who are 'confident' differ significantly from the two groups of students who are 'hardly confident' and ‘a little confident’. Therefore, all groups differ from one another. The results of the Kendall's tau-b and Spearman's rho indicate that there is a statistically significant correlation between the students' confidence level and the means of self-assessment scores. In conclusion, the subjects having high self-assessment scores are more likely to have high confidence in using English and vice versa.

This study aimed to determine the perceptions and motivations of EFL students at SMP St. Ignatius Medan who apply blended learning in English subjects. The collected data were analyzed descriptively and qualitatively. The research findings show that EFL students have a positive perception of the use of blended learning in learning English and an increase in their learning motivation after using this method. This is because 72.9% of students are more comfortable using technology in the classroom, and 53.1% of students have a more interactive and interesting learning experience, even though student feedback is considered to be of lower quality; namely, from the data obtained, only 24% of students easily get feedback from teachers. The main problem for students is that 57.3% of them feel there is a lack of social interaction with classmates and teachers, and 71.9% have problems with poor internet connectivity, which causes them to not fully understand learning material. Student motivation in blended learning is influenced by two factors: the first is the intrinsic factor, where it was found that 57.3% of students felt learning English with blended learning was more interesting and fun. Student motivation related to the extrinsic factor; researchers found that 67.7% of students felt motivated to learn English with blended learning because they could access learning materials more flexibly. From these data, it was found that the impact of learning English using blended learning was due to the intrinsic factor: 94.8% of students felt that English was important for their future, and 84.4% of students felt that learning English was a necessity. from the extrinsic impact factor of blended learning with students' motivation to learn English, 86.4% of students feel that they will get a better job in the future. The research findings show that the use of blended learning can be applied more broadly and effectively in increasing the learning motivation and English skills of EFL students.

In the 21st century, students must master digital literacy. During the COVID-19 pandemic, many students could still not use online learning support facilities. It has an impact on decreasing reading competence and student engagement. This study aims to determine the effect of peer assessment-based digital literacy on reading competency and engagement of EFL students during the post-COVID-19 pandemic. This research utilized a quasi-experimental design. A total of 10 groups of students participated in this study. Two groups of students were selected using cluster random sampling; 30 participants were assigned to the experiment, and 30 were assigned to the control group. The data instruments used two post-tests and a questionnaire. The dataset was statistically analyzed using One-way ANOVA and MANOVA, with assistance from SPSS version 26.0. The findings of this study implied that there was a significant effect of peer assessment-based digital literacy on the reading competency of EFL students; and there was a significant effect of peer assessment-based digital literacy on the EFL students' engagement. There is a simultaneous effect of peer assessment-based digital literacy on EFL students' reading competency and engagement. This study found that peer assessment-based digital literacy is one of the best ways to improve EFL students' reading competency and engagement. It is suggested that more research be done on this topic, not just on how well students read, but also on how motivated students are to learn, how students and parents see each other, how involved parents are, and what factors help and hinder students.

This study employed an action research framework to investigate the impact of translanguaging on listening comprehension and learning motivation among Japanese university‐level EFL students. The participants first listened to an English passage, discussed it in English‐only groups, and wrote English comprehension summaries. A translanguaging intervention was then introduced, whereby students were encouraged to leverage their full linguistic repertoires to enhance listening comprehension and bridge knowledge gaps in collaboration with their pairs. The participants listened to another English passage and engaged in group discussions using translanguaging before producing individual summaries in English. A comparison of the two comprehension summaries showed that participants scored significantly higher on theirs when they used translanguaging during discussions, producing longer summaries that included more main idea units with greater accuracy than when limited to English‐only use. Post‐task interviews further revealed that translanguaging supported both extrinsic and intrinsic motivation by promoting learner autonomy and validating students’ linguistic identities. Notably, several participants reported a shift in their attitudes, illustrating how a well‐structured translanguaging intervention can effectively reshape beliefs about language learning.