South Asian Languages Research Articles

Recognition and Classification of Handwritten Urdu Numerals Using Deep Learning Techniques

Urdu is a complex language as it is an amalgam of many South Asian and East Asian languages; hence, its character recognition is a huge and difficult task. It is a bidirectional language with its numerals written from left to right while script is written in opposite direction which induces complexities in the recognition process. This paper presents the recognition and classification of a novel Urdu numeral dataset using convolutional neural network (CNN) and its variants. We propose custom CNN model to extract features which are used by Softmax activation function and support vector machine (SVM) classifier. We compare it with GoogLeNet and the residual network (ResNet) in terms of performance. Our proposed CNN gives an accuracy of 98.41% with the Softmax classifier and 99.0% with the SVM classifier. For GoogLeNet, we achieve an accuracy of 95.61% and 96.4% on ResNet. Moreover, we develop datasets for handwritten Urdu numbers and numbers of Pakistani currency to incorporate real-life problems. Our models achieve best accuracies as compared to previous models in the literature for optical character recognition (OCR).

Hindu: A History

AbstractThis article provides a textured history of the multivalent term “hindu” over 2,500 years, with the goal of productively unsettling what we think we know. “Hindu” is a ubiquitous word in modern times, used by scholars and practitioners in dozens of languages to denote members of a religious tradition. But the religious meaning of “hindu” and its common use are quite new. Here I trace the layered history of “hindu,” part of an array of shifting identities in early and medieval India. In so doing, I draw upon an archive of primary sources—in Old Persian, New Persian, Sanskrit, Prakrit, Hindi, Marathi, Bengali, and more—that offers the kind of multilingual story needed to understand a term that has long cut across languages in South Asia. Also, I do not treat premodernity as a prelude but rather recognize it as the heart of this tale. So much of South Asian history—including over two thousand years of using the term “hindu”—has been misconstrued by those who focus only on British colonialism and later. We need a deeper consideration of South Asian pasts if we are to think more fruitfully about the terms and concepts that order our knowledge. Here, I offer one such contribution that marshals historical material on the multiform and fluid word “hindu” that can help us think more critically and precisely about this discursive category.

Ontology-Based News Linking for Semantic Temporal Queries

Daily newspapers publish a tremendous amount of information disseminated through the Internet. Freely available and easily accessible large online repositories are not indexed and are in an un-processable format. The major hindrance in developing and evaluating existing/new monolingual text in an image is that it is not linked and indexed. There is no method to reuse the online news images because of the unavailability of standardized benchmark corpora, especially for South Asian languages. The corpus is a vital resource for developing and evaluating text in an image to reuse local news systems in general and specifically for the Urdu language. Lack of indexing, primarily semantic indexing of the daily news items, makes news items impracticable for any querying. Moreover, the most straightforward search facility does not support these unindexed news resources. Our study addresses this gap by associating and marking the newspaper images with one of the widely spoken but under-resourced languages, i.e., Urdu. The present work proposed a method to build a benchmark corpus of news in image form by introducing a web crawler. The corpus is then semantically linked and annotated with daily news items. Two techniques are proposed for image annotation, free annotation and fixed cross examination annotation. The second technique got higher accuracy. Build news ontology in protégé using Ontology Web Language (OWL) language and indexed the annotations under it. The application is also built and linked with protégé so that the readers and journalists have an interface to query the news items directly. Similarly, news items linked together will provide complete coverage and bring together different opinions at a single location for readers to do the analysis themselves.

Improving User Intent Detection in Urdu Web Queries with Capsule Net Architectures

Detecting the communicative intent behind user queries is critically required by search engines to understand a user’s search goal and retrieve the desired results. Due to increased web searching in local languages, there is an emerging need to support the language understanding for languages other than English. This article presents a distinctive, capsule neural network architecture for intent detection from search queries in Urdu, a widely spoken South Asian language. The proposed two-tiered capsule network utilizes LSTM cells and an iterative routing mechanism between the capsules to effectively discriminate diversely expressed search intents. Since no Urdu queries dataset is available, a benchmark intent-annotated dataset of 11,751 queries was developed, incorporating 11 query domains and annotated with Broder’s intent taxonomy (i.e., navigational, transactional and informational intents). Through rigorous experimentation, the proposed model attained the state of the art accuracy of 91.12%, significantly improving upon several alternate classification techniques and strong baselines. An error analysis revealed systematic error patterns owing to a class imbalance and large lexical variability in Urdu web queries.

CORPURES: Benchmark corpus for urdu extractive summaries and experiments using supervised learning

• This study produces the Corpus for Urdu Extractive Summaries containing 161 documents with manually written extractive summaries from the news articles. • The supervised learning models can generate extractive summaries for the under-resourced Urdu language. • The preprocessing settings such as word segmentation, stemming & lemmatization improve the results. • The artificially generated datasets for extractive summaries do not improve the results. Text summarization is the process of shortening the text so that it conveys the key points. Several text summarization methods and benchmark corpora are available for languages like English. A significant hurdle in developing and evaluating existing or new text summarization methods is the unavailability of standardized benchmark corpora, especially for South Asian languages. Among other things, a reference corpus enables researchers to compare existing state-of-the-art methods. Our study addresses this gap by developing a benchmark corpus for one of the widely spoken yet under-resourced language Urdu. The reported corpus contains 161 documents with manually written extractive summaries from the newswire domain. We also perform several experiments on the corpus to show how it can be used to develop, evaluate, and compare text summarization systems using a supervised learning approach for the Urdu language. Our results show that the state of the art classifiers are good candidates for Urdu text summarization when supervised learning techniques are employed. Also, a radical word segmentation technique such as fixed-length segmentation outperforms all other settings (Senetnce Match F 1 = 57 % , ROUGE-2 F 1 = 64.4 % ). On the basic preprocessing of Urdu texts, we observe that tokenization of words on space is a reliable approach until the proper word segmentation tools for Urdu are mature enough. On word similarity features needed for supervised learning, it is observed that a radical stemming such as Ultra stemming with length (1 and 2) works better than the existing stemming and lemmatization tools for Urdu. Finally, the artificially generated datasets do not significantly improve results compared to the original data.

Deven M. Patel: Text to tradition: The Naiṣadhīyacarita and literary community in South Asia

Article Deven M. Patel. Text to tradition: the Naiṣadhīyacarita and literary community in South Asia was published on January 24, 2023 in the journal Journal of South Asian Languages and Linguistics (volume 0, issue 0).

Walter N. Hakala: Negotiating languages: Urdu, Hindi, and the definition of modern South Asia

Article Walter N. Hakala, Negotiating languages: Urdu, Hindi, and the definition of modern South Asia was published on January 23, 2023 in the journal Journal of South Asian Languages and Linguistics (volume 0, issue 0).

Religious difference, colonial politics, and Grierson’s Linguistic Survey of India

ABSTRACT Grierson’s Linguistic Survey of India (1903–1928) is one of the most complete sources on South Asian languages. It has influenced all subsequent studies of the language situation in India. However, there are indications in the Survey’s volumes, in its unpublished files, and in Grierson’s correspondence, that extra-linguistic considerations affected his approach to some Indian languages. Drawing on these sources, this essay focuses on Panjabi, Siraiki, Assamese, and Hindi-Urdu. It shows how factors stemming from Grierson’s views on religious difference and on language as a basis for nationality, as well as colonial politics of governance, may have influenced his characterisations of these languages. However, this does not invalidate the Survey, which is not straightforwardly ‘colonial’. Moreover, each of these languages is also described using linguistic argumentation, as reflected in the LSI’s skeletal grammars and its focus on dialectal variation. As such, we have to work with this tension in the LSI, without trying to resolve it either by rejecting the Survey in toto because of the instances of politics affecting its analyses, or by accepting it wholesale while ignoring the extra-linguistic considerations which influenced how it characterised some Indian languages.

“Language and the Periphery” Response to Book Forum on Insurgent Imaginations: World Literature and the Periphery

AbstractAuritro Majumder is Associate Professor of English at University of Houston. He is the author of Insurgent Imaginations: World Literature and the Periphery (Cambridge University Press, 2021) and currently chair of the South Asian and Diasporic Languages, Literatures and Cultures forum of the Modern Language Association.

Investment in Learning English: A Case Study of Chinese LOTE Learners

With the increasing global status of China’s economy, languages other than English education (LOTE) in China has experienced an unprecedented expansion size. The importance of offering language programs is particularly relevant and geographically significant in China’s border provinces. Yunnan, China’s Southwest border province, has been discursively positioned as an ideal space for cultivating South Asian and Southeast Asian language learners for China’s cross-cultural communication. However, while majoring in LOTE, Chinese students have to prove their proficiency in English both for academic attainment and employment prospect. This study examines how English is ideologically embedded in the learning process of LOTE learners in the context of China’s socioeconomic transformation and regional integration with its neighbouring countries. Adopting the concept of language, investment and ideology (Darvin & Norton, 2015), the study explores the English learning experiences of Chinese undergraduates majoring in LOTE in a Chinese border university in Yunnan. Findings reveal how LOTE learners attach great importance on English and see English as a symbolic capital which can be translated into academic credits and facilite their future educational and social mobility. Findings also demonstrate that LOTE learners take advantage of their learning strategies in English and transfer their learning strategies into their LOTE learning process. Based on the findings of the study, it is argued that LOTE students can be empowered to enhance their language proficiency by LOTE in the process of investment learning English and their investment can help them acquire additional multilingual ability and multiple identities. The study can prove pedagogical implications for curriculum of LOTE.

Sentiment analysis techniques, challenges, and opportunities: Urdu language-based analytical study.

Sentiment analysis in research involves the processing and analysis of sentiments from textual data. The sentiment analysis for high resource languages such as English and French has been carried out effectively in the past. However, its applications are comparatively few for resource-poor languages due to a lack of textual resources. This systematic literature explores different aspects of Urdu-based sentiment analysis, a classic case of poor resource language. While Urdu is a South Asian language understood by one hundred and sixty-nine million people across the planet. There are various shortcomings in the literature, including limitation of large corpora, language parsers, and lack of pre-trained machine learning models that result in poor performance. This article has analyzed and evaluated studies addressing machine learning-based Urdu sentiment analysis. After searching and filtering, forty articles have been inspected. Research objectives have been proposed that lead to research questions. Our searches were organized in digital repositories after selecting and screening relevant studies. Data was extracted from these studies. Our work on the existing literature reflects that sentiment classification performance can be improved by overcoming the challenges such as word sense disambiguation and massive datasets. Furthermore, Urdu-based language constructs, including language parsers and emoticons, context-level sentiment analysis techniques, pre-processing methods, and lexical resources, can also be improved.

Empire of Eloquence: The Classical Rhetorical Tradition in Colonial Latin America and the Iberian World

Empire of Eloquence: The Classical Rhetorical Tradition in Colonial Latin America and the Iberian World

South Asian Language Practices: Mother Tongue, Medium, and Media

Scholars such as Murray Emeneau and John Gumperz made India prominent in the development of sociolinguistics as a field of study through their simultaneous attention to difference and cohesiveness. Later, scholars stressed the ideological mediation of practice, especially the importance of colonial constructions that continue to be relevant in the postcolonial period. Work on specific notions such as mother tongue and medium of instruction, and the salience of English, led scholars to provide insights into multilingual practices in Bangladesh, Nepal, Pakistan, and Sri Lanka. Finally, a vast scholarship on an array of older and newer media forms ranging from early print publications to social media has posed questions about the possibilities of representation and participation. Ethnographic approaches to digital media that focus on the complex dynamics between ideologies and practices have put South Asia at the forefront of studies of communication.

Terrorism and Poverty against Child Right in South Asia: A Rhetorical Analysis of Speeches of Yousafzai and Satyarthi

This paper analyses the rhetoric and the content of the speeches on the Nobel Prize award ceremony delivered by South Asian Nobel Prize laureates Malala Yousafzai and Kailash Satyarthi. It is the argument that the speakers have exposed in international forum concerning child right and education. Their speeches encompass the South Asian philosophy, language discourse, speech and tone, Hindu and Islamic scripture’s chants and the problems of terrorism and poverty. It aims to evaluate the rhetorical elements comparatively followed in the speeches to argue and convince the audience. Terrorism, poverty and trafficking are the main problems against child right and education in South Asian countries against which they struggled and raised voice and they deserved the award. As for method, the article adopts the audio speech in YouTube and written speech of both speakers in Google. Various internet sources of rhetorical analysis are taken as secondary resources to study and analyse their speeches. The speeches are analysed from the perspective of ethos, logos and pathos which are the main rhetorical elements in public speaking as discussed by Aristotle. Satyarthi’s speech is found more inclusive and strong than of Yousafzai’s in terms of coverage and content. Both speakers are aware of using rhetorical skills while introducing South Asian problem in international forum.

Named Entity Recognition Using Conditional Random Fields

Named entity recognition (NER) is an important task in natural language processing, as it is widely featured as a key information extraction sub-task with numerous application areas. A plethora of attempts was made for NER detection in Western and Asian languages. However, little effort has been made to develop techniques for the Urdu language, which is a prominent South Asian language with hundreds of millions of speakers across the globe. NER in Urdu is considered a hard problem owing to several reasons, including the paucity of large, annotated datasets; an inaccurate tokenizer; and the absence of capitalization in the Urdu language. To this end, this study proposed a conditional-random-field-based technique with both language-dependent and language-independent features, such as part-of-speech tags and context windows of words, respectively. As a second contribution, we developed an Urdu NER dataset (UNER-I) in which a large number of NE types were manually annotated. To evaluate the effectiveness of the proposed approach, as well as the usefulness of the dataset, experiments were performed using the dataset we developed and an existing dataset. The results of the experiments showed that our proposed technique outperformed the baseline technique for both datasets by improving the F1 scores by 1.5% to 3%. Furthermore, the results demonstrated that the enhanced dataset was useful for learning and prediction in a supervised learning approach.

Values and beliefs about urban forests from diverse urban contexts and populations in the Greater Toronto area

Cities around the world are diverse. People’s perceptions of urban forests may vary according to urban contexts and people’s diverse identities. A better understanding of these diverse perceptions is critical to support stewardship initiatives, inform urban tree decisions, and guide community engagement, among other key management and governance processes in urban forestry. This study examines the values and beliefs that diverse people living in a variety of urban contexts associate with urban trees. Using an urban gradient approach, 2009 responses were collected through an online panel survey conducted in the Greater Toronto Area (GTA) to understand variations in values and beliefs of urban forests across municipal types (inner core, suburban, peri-urban, and regional cities). The GTA has an ethnoculturally diverse population, allowing us to also explore perceptions held by people with various identities. The study uses analytical techniques, such as means differences and linear regression models, to segment the diverse meanings people attribute to urban trees by municipal types and people’s identities. The results show that, while people value the urban forest very highly, people’s value orientations vary depending on their context and identities. For example, respondents who speak European or South Asian languages tend to hold value orientations related to cultural, social, and identity issues. Respondents living in inner municipal types tend to hold values related to identity issues, whereas respondents living in outer municipal types and regional cities tend to hold values related to natural issues. The results also show that people’s beliefs about urban forests are primarily positive. Urban forests managers and stewardship initiatives should recognize the diverse set of values and beliefs that people associate with urban forests, incorporating these perceptions into policy and programs. Additionally, since specific value and belief orientations may depend on personal identity and urban contexts, tailored messaging may also help generate support for policy or stewardship initiatives across different communities and in varied urban contexts.

Towards a Typology of Negation in South Asian Languages

The present study is a preliminary typology of negation in South Asian languages, based on a database of 25 structural features for 39 languages from three language families and two language isolates. The documented features include the form of the negative marker, the relation of the negative construction to the corresponding affirmative form, whether there are different negative constructions used in different TAM categories, and whether these constructions are symmetric or asymmetric. This study also provides a first analysis of the distribution of these different negative-marking strategies throughout the subcontinent and suggests that a combination of both family bias and areal pressure are needed to account for many of the observed distributions. In some cases of language contact, the data also allows us to determine with some certainty the type of contact situation which has led to the negative-marking patterns documented in the database.

Multi-label emotion classification of Urdu tweets.

Urdu is a widely used language in South Asia and worldwide. While there are similar datasets available in English, we created the first multi-label emotion dataset consisting of 6,043 tweets and six basic emotions in the Urdu Nastalíq script. A multi-label (ML) classification approach was adopted to detect emotions from Urdu. The morphological and syntactic structure of Urdu makes it a challenging problem for multi-label emotion detection. In this paper, we build a set of baseline classifiers such as machine learning algorithms (Random forest (RF), Decision tree (J48), Sequential minimal optimization (SMO), AdaBoostM1, and Bagging), deep-learning algorithms (Convolutional Neural Networks (1D-CNN), Long short-term memory (LSTM), and LSTM with CNN features) and transformer-based baseline (BERT). We used a combination of text representations: stylometric-based features, pre-trained word embedding, word-based n-grams, and character-based n-grams. The paper highlights the annotation guidelines, dataset characteristics and insights into different methodologies used for Urdu based emotion classification. We present our best results using micro-averaged F1, macro-averaged F1, accuracy, Hamming loss (HL) and exact match (EM) for all tested methods.

THE FORM AND FUNCTION OF REDUPLICATION IN ENGLISH AND URDU LANGUAGES: A CROSS LINGUIST STUDY

The main concentration of this work is to look at the forms and functions of morphological doubling in syntax and it aims at to look at the functions and roles of double questions morphemes of Urdu and to make comparative analysis of Urdu and English languages. There are different types of doublings in English language e.g. adverbs doublings, semantic doubling but question doubling does not exist in English language because it is the unique feature of Urdu language only. Qualitative analysis with purposive sampling approach, with the framework of Lexical Functional grammar (LFG).through the concept of F-Structure from (LFG) theory, double morphemes will be analyzed. For this purpose, data will be collected through document analysis, standard books, articles, published researches, dictionaries and data from common users. 100 sentences, consisting of double morpheme, will be selected and same were put into F-Structure to explore their lexical structure to show the semantic functions in both languages. Simultaneously, the sentences were also analyzed syntactically in order to find out their appropriateness. The study may provide a good level of confidence to develop Urdu National Language processing (UNLP) and enlighten the phenomena that morphological doubling in Urdu used for multiple semantic functions to explore human notions and feelings as compare to any other South Asian Language because Urdu is the only language with per gram. and study will also explore the functional and dynamic aspects of LFG. Keywords: Reduplication, LFG, syntax, Double interrogative morphemes.

A Fuzzy System for Identifying Partial Reduplication

Reduplication is a common feature more or less in almost all languages. It is a linguistic process that has been studied since back in both a morphological as well as a phonological process. Literary reduplication is the repetition of tokens in various forms like morpheme, phrase, word, etc. In most cases, it affects the syntactic or/and semantic meaning of the original words. But in several cases, it is not the exact reduplication rather partial reduplication. In the exact or total reduplication, it exactly reiterates a word or a phrase (e.g. fifty-fifty, bye-bye in English) while in partial reduplication reiteration partially (e.g. flip-flop in English). The morphological structure of a total or partial reduplication is relatively simple for English-like languages, but it is complex in the case of African, Austronesian, or South-Asian languages. Earlier researchers tried to address this problem using the various heuristic approaches. This paper presented a set of fuzzy-based approach to deal with partial reduplication in the Bengali language.