An Approach for Pronunciation Classification of Classical Arabic Phonemes Using Deep Learning

A mispronunciation of Arabic short vowels can change the meaning of a complete sentence. For this reason, both the students and teachers of Classical Arabic (CA) are required extra practice for correcting students’ pronunciation of Arabic short vowels. That makes the teaching and learning task cumbersome for both parties. An intelligent process of students’ evaluation can make learning and teaching easier for both students and teachers. Given that online learning has become a norm these days, modern learning requires assessment by virtual teachers. In our case, the task is about recognizing the exact pronunciation of Arabic alphabets according to the standards. A major challenge in the recognition of precise pronunciation of Arabic alphabets is the correct identification of a large number of short vowels, which cannot be dealt with using traditional statistical audio processing techniques and machine learning models. Therefore, we developed a model that classifies Arabic short vowels using Deep Neural Networks (DNN). The model is constructed from scratch by: (i) collecting a new audio dataset, (ii) developing a neural network architecture, and (iii) optimizing and fine-tuning the developed model through several iterations to achieve high classification accuracy. Given a set of unseen audio samples of uttered short vowels, our proposed model has reached the testing accuracy of 95.77%. We can say that our results can be used by the experts and researchers for building better intelligent learning support systems in Arabic speech processing.

Interesting Pausal Forms in the Speech of Muslims and Christians in Kufᵘr-Kanna

The aim of this paper is to describe the system of vocalic variants in pause pertaining to speakers of Arabic in Kufᵘr-Kanna (AKK) and in this regard to determine the features that characterize the AKK. As in Nazareth, the incidence of pauses in AKK varies and depends on the content, the listener and the speaker’s intentions. In AKK I detected pausal forms in the speech of middle-aged and elderly Muslims and elderly Christians. In addition to changes in consonants and vowel quality in their speech, in pausal position final syllables also undergo other modifications as compared to the contextual forms. Unlike in Nazareth, four further types were identified in AKK: (1) lengthening of short vowels in final position: ‑Cv > ‑Cv̄#, -CvC > -Cv̄C#; lengthening of normal and anaptyctic short vowels in final closed syllables: -CvC#; (2) devoicing of voiced consonants in word-final position; (3) glottalization after con­sonants and vowels in word-final position; and (4) aspiration: addition of (h) in pausal position where the word ends in long vowels. Key words: Arabic dialects – Pausal forms – Syllables – Long vowels – Short vowels – Christians and Muslims.

Lingvistisko īpatnību paralēles kursiskajās un dziļajās tāmnieku izloksnēs

The present paper aims to show the most significant phonetic and morphological features belonging both to the deep Tamian (tāmnieku) subdialects of the Livonic dialect and the Curonian (kursiskās) subdialects of the Middle dialect. The main focus is on the forms used in the deep Tamian subdialects of Pope, Ziras, Piltene, Zlēkas, Ance, Dundaga, and Ugāle, based on the materials acquired by the author during the period 2008–2019, and compared with the data of dialect archive of the Latvian Language Institute of the University of Latvia (collected in the 50s and 60s of the 20th century). In phonetics, a typical feature is the use of broad e, ē, especially in infinitives, e. g., celt ‘to raise, to build’, sēt ‘to sow’, meklēt (Curonian subd.), meklet (Tamian subd.) ‘to search’, the lengthening or diphthongization of short vowels a, e, i, u before consonant r, e. g., dârps ‘work’, ʒērt ‘to drink’, zîrks // ziêrks ‘horse’, bũrt // buõrt ‘to conjure’, palatal consonant ŗ, e. g., kaŗš ‘war’, gaŗš ‘long’, jũŗ ‘sea’, also the use of vowel u before the consonant v or b, e. g., zuve (Curonian subd.), zȗu // zȗi (Tamian subd.) ‘fish’, suvēnc (Curonian subd.), suvēnc // suvans (Tamian subd.) ‘pig’, dubans // dubenc ‘bottom, seat’, dui // dȗv (Tamian subd.) ‘two’, the loss of sounds in different positions, e. g., cilēks (Curonian subd.), cileks (Tamian subd.) ‘human’, ciris (Curonian subd.), cȋrs (Tamian subd.) ‘axe’, dures // duõres (Curonian subd.), dȗrs // dûrs // duôrs (Tamian subd.) ‘the doors’, tēs ‘father’, vẽš ‘wind’, rupš ‘rude’, and assimilation of -ln- to -ll-, e. g., mells ‘black’, vells ‘the devil’, villa (Curonian subd.), vill (Tamian subd.) ‘wool’, pills ‘full’. In morphology, a characteristic feature is the use of ē-stem substantives, e. g., klẽte (Curonian subd.), klẽt (Tamian subd.) ‘barn’, pirte // piêrte (Curonian subd.), pirt // piêrt (Tamian subd.) ‘bathhouse’, sirde // siêrde (Curonian subd.), sȋrd (Tamian subd.) ‘heart’, also the use of archaic plural dative endings -iems or -ems, e. g., zirgiêms // zîrgiêms (Curonian subd.), zirgems // zîrgems (Tamian subd.) ‘for horses’, -ams: mãsams ‘for sisters’, -ems or -ems, e. g., guõvems (Curonian subd.), guõvems (Tamian subd.) ‘for cows’, and the use of prefix / preposition az or āz, e. g., âzvakãr // azvakar (Curonian subd.), azvakar // aizvakar // aîzvakar (Tamian subd.) ‘the day before yesterday’, âz // az kalna (Curonian subd.), az // aiz // aîz kal:n (Tamian subd.) ‘behind the hill’, also prefix or preposition ūz, e. g., ũzlikt (Curonian subd.), ûzlikt (Tamian subd.) ‘to put on’, ûz galda (Curonian subd.), ûz gal:d (Tamian subd.) ‘on the table’, also a typical peculiarity of the Curonian and Tamian subdialects (but not heard nowadays) is the use of pronouns tau, sau ‘for yourself, for myself’.

Syllable weight, vowel length and focus in Lebanese Arabic

Research on a variety of languages has shown that vowel duration is influenced by phonological vowel length as well as syllable structure (e.g., Maddieson, 1997). Further, the phonological concept of a mora has been shown to relate to phonetic measurements of duration (Cohn, 2003; Hubbard, 1993; Port, Dalby, & O'Dell, 1987). In Levantine Arabic, non-final closed syllables that contain a long vowel have been described as partaking in mora-sharing (Broselow, Chen, & Huffman, 1997; Khattab & Al-Tamimi, 2014). The current investigation examines the effect of vowel length and syllable structure on vowel duration, as well as how this interacts with durational effects of prosodic focus. Disyllabic words with initial, stressed syllables that were either open or closed and contained either a long or a short vowel wereexamined when non-focused and in contrastive focus. Contrastive focus was associated with longer words and syllables but not vowels. Short vowels were shorter when in a syllable closed by a singleton but not by a geminate consonant, while long vowels were not shortened before coda singletons. An analysis is proposed whereby long vowels followed by an intervocalic consonant cluster are parsed as open syllables, with the first consonant forming a semisyllable (Kiparsky, 2003), while long vowels followed by geminate consonants partake in mora-sharing (Broselow, Huffman, Chen, & Hsieh, 1995). The results also indicate compensatory shortening for short vowels followed by a singleton coda.

Framework of diacritic segmentation for Arabic handwritten document

<span lang="EN-US">In <span>recent Arabic standard language and Arabic dialectal texts, diacritics and short vowels are absent. There are some exceptions have been made for the Arabic beginner learner scripts, religious texts and as well as a significant political text. In addition, the text without diacritics is considered ambiguous due to numerous words with different diacritic marks seem identical. However, this paper we present a framework for segmenting diacritics from Arabic handwritten document by using region-based segmentation technique. Since Arabic handwritten and Mushaf Al-Quran contain many diacritical marks. Hence, the diacritics must be properly extracted from Arabic handwritten document to avoid losing some good features. Furthermore, the proposed framework is devised specifically to segment diacritics from Arabic handwritten image, thus there will be no feature extraction, feature selection, and classification processes included. Besides, we will present the methodology that is used to fulfil the objectives of this paper. The pre-processing phases will be explained and more specifically segmentation phase for segmenting diacritics which is the phase we concentrate more in this article. Lastly, we will identify the proposed technique region-based segmentation to facilitate our development throughout the experimental process.</span></span>

Interaction of Lithuanian and Belarusian Dialects in the Surroundings of Radun

The research focuses on the interaction of the Lithuanian and Belarusian languages in thesurroundings of Radun (Belarus, Voranava district, Lith. Rodūnia). Employing the language changemethodology, the paper presents a detailed analysis of the most prominent cases of phonetic interference oftwo – old and middle – generations of residents in both languages spoken by them.The data show that the oldest residents are expressing both languages using the Lithuanian articulatory base. Over time, the characteristic Lithuanian features begin to level: varies the pronunciation of a; in open endings, the long tense vowels i, ы are usually pronounced as é; unstressed vowels are reduced inconsistently, etc.As Lithuanian is for the middle-generation speakers, it already demonstrates essential changes, whichoccurred because they started expressing their Lithuanian using the Belarusian articulatory base. First of all, they usually lengthen stressed short vowels; they do no longer pronounce unstressed long vowels; they donot distinguish between acute – or circumflex-accented mixed i, u + R type diphthongs; the distinction between the falling and rising diphthongs ai, au, is made based on phonetic, not accentual, factors; the palatalised dental consonants t, d are replaced by the palatalised velar consonants k, g.

Valodniecības pētījumu nozīmīgums logopēda profesijā

The study has been carried out within the subproject No. 8 “Latvian Language Acquisition” framework of the National Research Programme “Latvian Language”. The aim of the study is to research and describe the importance of linguistic research results to speech and language therapists. In the course of studying children’s speech in 2019 and 2020 in different cities of Latvia, it was found that children of pre-school age, regardless of their mother tongue, have widespread sound pronunciation disorders, e.g. elision of the sound r (bukāns, tīs) or substitution of the sound with another sound (klāsas, klūze, gjiež). However, minority children speaking in Latvian have additional difficulties pronouncing the sounds that are not present in their native Russian. Frequently, long vowels are pronounced as short vowels, e.g. when asked what it is/they are? (ābols, krūze, šķēres, grāmata, burkāns) and what it does/they do? (ēd) the children answer abols, kruze, šķeres, gramata, burkans ed, etc., because, in their native language, there are only short vowels. The substitution of broad vowel e [ae] with the narrow e, e.g. ed [ed], is caused by the same reason. Pronunciation of the diphthong o [uo] as a long vowel has also been observed, e.g. pōgas (pogas). When children are not sure of the proper ending of a word, they do not pronounce it, e.g. abol (ābols), putn (putni), žiraf (žirafe). The speech studies of minority children also showed a large number of grammatical and vocabulary errors, which affect the development of their language and may, in the future, lead to difficulties and disruptions in the learning of the language, which, in more severe cases, lead to a disruption in the learning of subjects at school. The study results showed that in Latvia, the bilingualism of the minority pre-school children is relative. In general, only children whose families are bilingual or who attend the pre-school education groups with Latvian as the communication language speak well both their native language and Latvian. Observations during the National Research Programme show that speech therapists increasingly face the demand for minority children to learn Latvian as the second language; however, speech therapists do not have information about implementing corrections in such cases. Therefore, research on the characteristics of bilingualism and results of language studies are useful for speech therapists in diagnosing, correcting, and preventing disorders. The speaking and language skills of many minority children of pre-school age are still insufficient to continue successful learning in Latvian or bilingually. There are many phonetical and grammatical errors in their speech, while a poor vocabulary interferes with expressing thoughts and engaging in conversation. Both teachers and speech therapists need modern theoretical and practical training to raise their qualifications and implement a common understanding of the problems. The lessons learned from this project are incorporated into the bachelor’s theses of Marta Jakušina, Valērija Krasovska, and Alise Valtere.

A Historical Reconstruction of Some Pronominal Suffixes in Modern Dialectal Arabic

The morphology of the pronominal suffixes in dialectal Arabic are of particular interest for scholars of the history of Arabic for two main reasons. First, multiple dialects attest suffixes that, from a comparative perspective, apparently retain final short vowels. The second and more complicated issue concerns the vowels which precede the suffixes in the dialects, which are thought to either have been case inflecting or epenthetic. In this paper, I take up Jean Cantineau’s “embarrassing question” of how to account for the development of the vowels of the pronominal suffixes. Based on data from dialectal tanwīn in modern dialects, and attestations from pre-modern texts as well, I will argue that the pre-suffix vowels did originate in case inflecting vowels, but that no historical model heretofore proposed can satisfactorily account for how the various dialectal forms might have arisen. I identify two major historical developments and propose models for each. First, I suggest that dialects in which the pre-suffixal vowels harmonized with the suffix vowels developed via a process of harmonization across morpheme boundaries before the loss of final short vowels. For dialects in which one vowel is generalized, I argue that a post-stress neutralization took place, which led to a single vowel both before suffixes and tanwīn as well. Finally, I rely on evidence from the behavior of the suffixes to argue that the final vowel of the 3fs suffix was originally long, but that those of the 3ms, 2ms, and 2fs were most likely short.

(Mis) perception of consonant clusters and short vowels in English as a foreign language

Research addressing second language (L2) speech is expanding. Studies increasingly demonstrate that a learner’s first language (L1) filters the L2 input, resulting in learners misperceiving what they have heard. This L1 filter can result in learners perceiving sounds not actually present in the input. We report on a study which explored English consonant clusters and short, unstressed vowel perception of 70 Arabic-, Mandarin-, Spanish-speaking foreign language learners and 19 native English speakers. These are the vowels which speakers from two of the L1s typically insert in their production of English to break up L1-disallowed consonant clusters and the schwa which is documented to cause both perception and production problems. Results show that participants misperceive stimuli containing consonant clusters and counterparts where clusters are broken up by epenthetic/prothetic elements. In line with Sakai, Mari & Colleen Moorman 2018. We call for the inclusion of such findings on perception in pedagogical advice on pronunciation.

Patterns of Consonant Clusters in Word Initial, Medial, and Final Positions in Yemeni Arabic

Yemeni Arabic (YA) has a significant number of consonant clusters in word initial, medial, and final positions. However, their frequency of usage is not uniform. This study aims to investigate the patterns of consonant clusters in word initial, medial, and final positions in YA and also to find out the most and least frequent clusters in terms of their percentage. Qualitative and quantitative methods were used in data analysis in this study. All the words were analyzed using Statistical Package for Social Sciences (SPSS). There are features of consonant clusters in Yemeni Arabic which differ from Modern Standard Arabic and some other Arabic dialects. In Yemeni Arabic, there are 29 consonants and 10 vowels, 5 long and 5 short vowels. The maximum number of onset cluster is three (e.g. /∫tsu:q/ “she will drive” while coda cluster is two (e.g. /satˤħ/ “roof”). Furthermore, the maximum number of medial clusters are also two (e.g. /muχ.lsˤu/ “sincere”. The analysis undertaken will throw light on the frequency and percentages of the occurrences of the consonant clusters on the basis of a word list, which is justified with the help of statistical support.