Phoneme Level Research Articles

An Error Analysis of Nonword Repetition Performance in Toddlers and Young Children: Theoretical Implications and Clinical Utility.

Purpose Nonword repetition has been cited as a measure of phonological working memory and continues to gain status as a clinical tool used to identify language impairment in school-age children. Less is known about nonword repetition skills in the toddler population. Method The current study presents a detailed analysis of errors by segmenting nonwords into word, syllable, and phoneme levels. Errors were also analyzed for type (e.g., addition, substitution, deletion). The Test of Early Nonword Repetition was used to measure performance in a sample of 36 typically developing children, aged 24-48 months. Clinical assessments including parent report, language sampling procedures, and standardized assessments were also administered. Results As a group, participants produced significantly more syllable errors compared to word-level errors; however, most errors were made at the phoneme level. Errors of addition were the least common error type, and no differences between substitutions or deletions were present for the entire sample. Toddlers (aged 2 years) produced more syllable-level errors compared to older children (aged 3 years). Substitution errors were positively correlated with performance on clinical measures of language, whereas deletion errors were negatively correlated with performance. Conclusion Nonword repetition performance patterns in young children may be associated with language delay or language impairment and have both clinical and theoretical relevance.

A Reading in the Bergstraße Approach to Making the Linguistic Atlas

Linguistic atlas is a science that records linguistic phenomena on geographical maps, in order to know their regional positions and characteristics. In light of this, the present study deals with a new science, which is geographical linguistics, that the ancient Arabs had significant contribution in it, to laying out its origins, by their research in the languages of the tribes and their dialects. As for the modern name and the completion of atlases, it is counted for Westerners. This field is increasing with the emergence of digital atlases, and the provision of a database.
 In this research, we will provide a reading of the German researcher Bergstraße's methodology in making an Arabic atlas, and explain the extent of his contribution to establishing a database that helps researchers to embody their achievements in this field.
 This researcher has studied the following language levels: the phonemic level, such as the form of pronouncing Kaaf, the level of formulas and pronouns, and the level of vocabulary. He used the descriptive approach, which is an approach based on recording linguistic real events.

Improving phonological awareness with Talking Tables in at-risk kindergarten readers

BackgroundFor young children experiencing phonological awareness (PA) difficulties, the need for early and targeted intervention to prevent reading disability is unequivocal. There are very few studies, however, on the efficacy of PA interventions delivered at school. AimsThis study examined the impact of an early PA intervention embedded within an oral language program designed for at-risk kindergartners. MethodsUsing a quasi-experimental pretest/posttest design, at-risk readers from four schools received either the 10-week intervention in small groups, three times a week for 30 min as a supplement to the regular classroom curriculum or served as controls not participating in the intervention and receiving the usual classroom instruction. ResultsChildren in the intervention group demonstrated a greater use of phonological awareness at posttest on overall composites of phonological processing, and on several individual accuracy and fluency measures targeting skills at the phoneme level. ConclusionsThe results add to accumulating evidence on the efficacy and effectiveness of teacher-delivered school-based early literacy interventions.

Coherence Between Brain Activation and Speech Envelope at Word and Sentence Levels Showed Age-Related Differences in Low Frequency Bands.

Speech perception is dynamic and shows changes across development. In parallel, functional differences in brain development over time have been well documented and these differences may interact with changes in speech perception during infancy and childhood. Further, there is evidence that the two hemispheres contribute unequally to speech segmentation at the sentence and phonemic levels. To disentangle those contributions, we studied the cortical tracking of various sized units of speech that are crucial for spoken language processing in children (4.7-9.3 years old, N = 34) and adults (N = 19). We measured participants' magnetoencephalogram (MEG) responses to syllables, words, and sentences, calculated the coherence between the speech signal and MEG responses at the level of words and sentences, and further examined auditory evoked responses to syllables. Age-related differences were found for coherence values at the delta and theta frequency bands. Both frequency bands showed an effect of stimulus type, although this was attributed to the length of the stimulus and not the linguistic unit size. There was no difference between hemispheres at the source level either in coherence values for word or sentence processing or in evoked response to syllables. Results highlight the importance of the lower frequencies for speech tracking in the brain across different lexical units. Further, stimulus length affects the speech-brain associations suggesting methodological approaches should be selected carefully when studying speech envelope processing at the neural level. Speech tracking in the brain seems decoupled from more general maturation of the auditory cortex.

An Arabic Mispronunciation Detection System Based on the Frequency of Mistakes for Asian Speakers

Over the last few decades, the field of artificial intelligence and machine learning has evolved. Due to the advancement in these fields, much work has been done to assist language learning with the help of computers called Computer-Assisted Language Learning (CALL). Mispronunciation detection is one of the significant tasks of the CALL system. An efficient mispronunciation detection model has a positive impact on the life of second language learners by providing phoneme level feedback. In this paper, we introduce the phone grouping technique for mispronunciation detection that is based on mistakes probability. We consider mispronunciation detection as a classification problem, traditionally for this purpose, a separate classifier is trained for each phoneme mistake that requires a lot of memory and time. Instead of training a separate classifier, we group the phoneme based on their mistakes probability that helps in reducing the number of the classifiers to be trained and also saves memory and time. We use the Support Vector Machine (SVM) classifier and test the results on the Arabic dataset (28 Phonemes). The performance of our proposed method is evaluated by using accuracy. The results of the model are evaluated using the confusion matrix and gives an accuracy of 88%. Our approach outperforms the existing systems developed for Arabic phonemes in terms of accuracy and is also time/memory efficient.

Integral perception, but separate processing: The perceptual normalization of lexical tones and vowels

In tonal languages, speech variability arises in both lexical tone (i.e., suprasegmentally) and vowel quality (segmentally). Listeners can use surrounding speech context to overcome variability in both speech cues, a process known as extrinsic normalization. Although vowels are the main carriers of tones, it is still unknown whether the combined percept (lexical tone and vowel quality) is normalized integrally or in partly separate processes. Here we used electroencephalography (EEG) to investigate the time course of lexical tone normalization and vowel normalization to answer this question. Cantonese adults listened to synthesized three-syllable stimuli in which the identity of a target syllable — ambiguous between high vs. mid-tone (Tone condition) or between /o/ vs. /u/ (Vowel condition) — was dependent on either the tone range (Tone condition) or the formant range (Vowel condition) of the first two syllables. It was observed that the ambiguous tone was more often interpreted as a high-level tone when the context had a relatively low pitch than when it had a high pitch (Tone condition). Similarly, the ambiguous vowel was more often interpreted as /o/ when the context had a relatively low formant range than when it had a relatively high formant range (Vowel condition). These findings show the typical pattern of extrinsic tone and vowel normalization. Importantly, the EEG results of participants showing the contrastive normalization effect demonstrated that the effects of vowel normalization could already be observed within the N2 time window (190–350 ms), while the first reliable effect of lexical tone normalization on cortical processing was observable only from the P3 time window (220–500 ms) onwards. The ERP patterns demonstrate that the contrastive perceptual normalization of lexical tones and that of vowels occur at least in partially separate time windows. This suggests that the extrinsic normalization can operate at the level of phonemes and tonemes separately instead of operating on the whole syllable at once.

Effect of pitch enhancement in Punjabi children's speech recognition system under disparate acoustic conditions

In this work, a Punjabi children speech recognition system is developed under different acoustic matched and mismatched conditions. One major problem in children's speech recognition is the differences in the acoustic attributes of the children and adult speech signals, which leads to the poor recognition rate for the children's speech. This paper shows how pitch enhanced features extracted from the front-end feature extraction process plays an important role under mismatched acoustic conditions. After enhancing the pitch using the Cepstral analysis in the feature extraction process, the recognition rate of the children's speech recognition system using different age group datasets increases as compared to the normal acoustics features extracted using Mel Frequency Cepstral Coefficient (MFCC) feature extraction process. Kaldi toolkit is used for building the children's speech recognition models at different phoneme levels. Results show the improvement of 0.03% to 16.47% WER under different acoustic conditions using pitch enhanced features.

Consciência fonológica, nomeação automática rápida e leitura em adultos analfabetos funcionais

The study's purpose is to describe the cognitive profile of a sample of functionally illiterate individuals in reading, phonological awareness (PA) and rapid automatized naming (RAN), as well as to correlate the performance of these tasks. Moreover, it sought to understand how the performances in PA and RAN predict results in reading of words and pseudowords. 23 functionally illiterate adults were assessed for intelligence, reading, PA and RAN tasks. Participants showed difficulties in PA, performing poorly in tasks involving phoneme analysis and manipulation. In RAN, they found it easier to name alphanumeric items. Regarding reading, they presented difficulties in phonological processing, with errors caused by phonological exchange. Correlation analyses indicated that reading presented higher correlations with RAN than PA. Finally, a regression analysis indicated that performance in RAN can account for more than half of participants' reading results. In general, the profile of functionally illiterate adults presents impairment in the abilities assessed that are more evident in PA at phoneme levels as well as in naming of non-alphanumerical items. Moreover, the results in RAN explain those in word and pseudowords reading better than the results in PA.

On the revealing the emotional valence in communication by text

The interaction between humans and between humans and machines is often based on texts. The statistical analysis, presented here, provides one more reason to suppose that meaningful language communication integrates an emotional component involved at the phonemic level. The performed analysis is of written English language, presented by means words’ phonetic transcripts. The phonemic content of the texts is thought in layers, supposing that the phonological forms involved in the communication have the role of phonological markers assisting emotion transfer. The words are considered as composed of biphones – phonemic pairs consisting of one vowel and one consonant, in either order. Phonemic metadata for this sublexical representation was obtained from a vast corpus of emotionally evaluated texts and submitted to statistical analysis. The result of the applied principle component regression showed that the phonemic content is very strongly related (r = 0.96) with the ratings of emotional valence (positive-negative emotion), provided by readers. The analysis showed the big importance of the first biphone in the words, in alliance with results for single phonemes reported in recent studies. The result shows that the phonemic content presented as a stream of biphones is strongly related to the emotional impact that a text has on the reader. This leads to suppose that the mechanisms that have shaped the phonological forms to communicate conceptual meaning involve emotional valence too.

CAN AUDIO-VISUAL TRAINING EQUALLY AFFECT PHONEMIC AND PHONETIC CONTRASTS? AN EXAMPLE OF L2 FRICATIVE PRODUCTION

The paper explores the effect of audio-visual perceptual training on Serbian EFL learners’ production of novel phonemic and phonetic contrasts in L2, specifically focused on fricatives. Hence, the paper aims at discovering whether audio-visual training has equal effects at phonemic and phonetic levels, and also, whether the effect is the same at two different age/proficiency levels, 6th grade primary and 4th grade secondary school. In order to explore the phonemic level we concentrated on interdental fricatives, and for the phonetic level differences sibilant contrasts were included, following the predictions of the Perceptual Assimilation Model (Best 1994) and Speech Learning Model (Flege 1995). The testing for relevant acoustic information was per- formed prior to and immediately following the experimental period, when all the participants were recorded pronouncing a prepared sentence list containing target sounds. It consisted of measuring spectral moments, frication duration and comparison of spectrograms. The results of the audio-visual phonetic training proved especially beneficial for phonemic contrasts, i.e. interdental fricatives for both levels of age/proficiency, while sibilant contrasts showed insignificant progress. The age/proficiency level did not appear to be a significant predictor of the effect of audio-visual training. Along with the empirical results, the paper likewise presents pedagogical implications important for pronunciation teaching and highlights the significance of phonetic training in the Serbian EFL context in particular.

Contrastive Linguistic Study of South and North Korean Sign Language and Japanese Sign Language at the Level of Phoneme and Lexis

Contrastive Linguistic Study of South and North Korean Sign Language and Japanese Sign Language at the Level of Phoneme and Lexis

SpeakEasy: Browser-based automated multi-language intonation training

Learning to speak a second language involves a cycle of observation, mimicry, and feedback. A student observes a teacher, attempts to copy the teacher's performance, and then the teacher provides feedback on how the student performed. When the students' access to feedback is limited, so is their ability reinforce their learning. In this work a web based application, Speakeasy tools, in introduced to provide remote students with automated visual intonation feedback for multiple languages. For this study, participants are selected from the pool of Speakeasy users and their interactions with the application are observed over a set period of time. The application presents participants with native speaker examples generated via text-to-speech in the form of audio samples, fundamental frequency visualizations, and grapheme and phoneme level timelines. Participants are able to record and review an unlimited number of practice attempts which are processed using the same pipeline used for the native speaker examples. Using the Root-Mean-Square Error (RMSE) between native and participant fundamental frequencies over time, the practice attempt is assigned a score. Practice-attempt scores with respect to time spent using the application provide a metric for measuring a participant's progress. [Work supported by RPI Seed Grant and CISL.]

Phonological Awareness Skills in Children With Early and Late Cochlear Implantation: Effects of Task and Phonological Unit.

Purpose Phonological awareness (PA) skills are critical for spoken language acquisition and literacy. PA manifests in various skills that can be identified based on task performance and speech sound unit size. This study compared the PA skills of children with early cochlear implantation (E-CI), children with late cochlear implantation (L-CI), and children with typical hearing (TH) in relation to task and phonological unit. It also attempted to identify the significant predictors of PA skills in each CI and TH group. Method Twenty children with E-CI, 20 children with L-CI, and 20 children with TH participated in this study. PA skills were assessed using elision, blending, and segmenting tasks at both the syllabic and phonemic levels. Results The E-CI and L-CI groups performed significantly less well than the TH group on the elision and blending tasks at the syllabic level. However, the E-CI group performed at a similar level as the TH group in the segmenting tasks at both the syllabic and phonemic levels. The regression analysis identified age at implantation and receptive vocabulary scores as significant predictors of PA skills in children with CIs. Conclusions Although all the children with CIs had age-appropriate receptive vocabulary skills, the PA skills of both the E-CI and L-CI groups tended to lag behind those of the TH group in the elision and blending tasks at the syllabic level. Age at implantation and receptive vocabulary skills affected the development of PA skills in children with CIs.

The linguistic use of the group on the syntactic, morphological and phonemic levels

The linguistic use of the group on the syntactic, morphological and phonemic levels

Digital Masorah: Toward an Index of Orthographic and Morphological Variation at the Lexical Level

This article proposes a computational method to extract orthographic and morphological variation within the boundaries of individual documents that are part of the Hebrew Bible or the Dead Sea Scrolls. In particular, it presents “profiles” of document-internal variation at the lexical level, e.g., at the level of individual words. This study a) describes the methods used to create an index of variation that can be navigated and explored in order to find meaningful patterns, and b) explores the new avenues for research such an index opens up. Whereas earlier contributions have focused on the underlying rules that affect linguistic variation, the aim of the present contribution is to create a representative sample of that variation. This sample is not at the level of individual phonemes or morphemes, but at the lexical level as there are clear indications that lexical selection played a significant role in ancient linguistic practices.

PHONEMIC INTERFERENCE AND OVERREGULARIZATION IN THE /S/ AND /∫/ PHONEMES REALIZATION IN FRENCH

One of the problems in foreign language learning is interference, a rearrangement of patterns resulting from the presence of foreign elements in the language domain (Weinreich, 2010). This research shows how and why phonemic interference of /s/ and /∫/ phonemes occur from Indonesian and English although both phonemes exist in all three languages. Some interference begins from lexeme and then to phonemic level. Other faults are overregularization which is the application of regular grammatical patterns to irregular cases. This seems to support the Logical Problem of Language Acquisition which states that a student cannot correct his/her mistakes without explicit feedback from the linguistic environment (Pinker, 2004).The results of this research indicate that foreign language learning requires knowledge of non-structural elements that are outside of the language, not only following phonological, syntactic, morphological, or lexical rules (structural elements). For example, students' foreign language knowledge and cultural content in teaching materials.

PEMETAAN PENGGUNAAN BAHASA MELAYU BANGKA

This study aims to map the use of the Bangka Malay language. This research uses qualitative. The source of the data was observations, interviews with language users, and questionnaires focused on four districts, namely Bangka Regency, South Bangka Regency, Central Bangka Regency, West Bangka Regency, and one Madya City, Pangkalpinang. The reduction data for dialectological studies contain differences in phonological and lexical levels selected from all data obtained, except data in the form of phrases and sentences, while reduction data for comparative historical linguistic studies are collected from 200 basic Swadesh vocabularies (for lexicostatistic analysis), in which describe the word kin. The results of this study are as follows: a. The Malay language of Bangka has the same variation between user regions, especially those used in the City of Pangkalpinang, Sungailiat, Toboali, Koba, and Mentok; b. Some vocabularies are different between regions but are not significant because they occur only at the phonemic level such as the use of the words 'akar (root)' and 'aker', 'dingin (cold)' and 'dingen', 'dengar (hear)' and ‘denger'; c. Significant differences occur only in some of the vocabulary used by the Bangka community such as the word 'burn' which translates to 'menam', 'tembung', and 'tunu'; d. Vocabulary differences are generally found in remote areas such as Sadai, Air Gegas, and Kelapa.

الأخطاء فى تعليم مهارة الكلام وحلها عند اللغة المرحلية بالمستوى الفونيمات فوق القطعية للطلاب الفصل الثامن بمدرسة محمدية ٤ المتوسطة سورابايا

The difficulties in learning maharah kalam at the level of suprsegmental phonemes are found when students learn Arabic. Intonation errors, stress, and pausing. Therefore, this research was conducted to find out these mistakes and the solutions given by teacher in SMP Muhammadiyah 4 Surabaya. This research is a qualitative descriptive with the population of 8th grade student and the sampling is 8th grade C. Data collection techniques used in this research are using the method of observation, interview and questionnaire. Data analysis techniques used aredata reduction, data display, and data verification. The results showed: 1) errors in word stress were 49 errors; 2) errors in intonation were 30 errors; 3) errors at pauses were 55 errors. The solutions given by teacher with these mistakes were: 1) instructing students to repeat the dialogue to use the right intonation, stress, dan pause; 2) translating the text of theconversation and students writing the translation in their books, then the students will better understand the difference in meaning from the differences in intonation, stress, dan pause; 3) the teacher gives examples of the correct pronounciation of words according to the right intonation, stress, and pause.

Fine-grained statistical structure of speech.

In spite of its acoustic diversity, the speech signal presents statistical regularities that can be exploited by biological or artificial systems for efficient coding. Independent Component Analysis (ICA) revealed that on small time scales (∼ 10 ms), the overall structure of speech is well captured by a time-frequency representation whose frequency selectivity follows the same power law in the high frequency range 1–8 kHz as cochlear frequency selectivity in mammals. Variations in the power-law exponent, i.e. different time-frequency trade-offs, have been shown to provide additional adaptation to phonetic categories. Here, we adopt a parametric approach to investigate the variations of the exponent at a finer level of speech. The estimation procedure is based on a measure that reflects the sparsity of decompositions in a set of Gabor dictionaries whose atoms are Gaussian-modulated sinusoids. We examine the variations of the exponent associated with the best decomposition, first at the level of phonemes, then at an intra-phonemic level. We show that this analysis offers a rich interpretation of the fine-grained statistical structure of speech, and that the exponent values can be related to key acoustic properties. Two main results are: i) for plosives, the exponent is lowered by the release bursts, concealing higher values during the opening phases; ii) for vowels, the exponent is bound to formant bandwidths and decreases with the degree of acoustic radiation at the lips. This work further suggests that an efficient coding strategy is to reduce frequency selectivity with sound intensity level, congruent with the nonlinear behavior of cochlear filtering.

Improving Low-Resource Speech Recognition Based on Improved NN-HMM Structures

The performance of the ASR system is unsatisfactory in a low-resource environment. In this paper, we investigated the effectiveness of three approaches to improve the performance of the acoustic models in low-resource environments. They are Mono-and-triphone Learning, Soft One-hot Label and Feature Combinations. We applied these three methods to the network architecture and compared their results with baselines. Our proposal has achieved remarkable improvement in the task of mandarin speech recognition in the hybrid hidden Markov model - neural network approach on phoneme level. In order to verify the generalization ability of our proposed method, we conducted many comparative experiments on DNN, RNN, LSTM and other network structures. The experimental results show that our method is applicable to almost all currently widely used network structures. Compared to baselines, our proposals achieved an average relative Character Error Rate (CER) reduction of 8.0%. In our experiments, the size of training data is ~10 hours, and we did not use data augmentation or transfer learning methods, which means that we did not use any additional data.