Corpus Research Center

The Role of Islamic Financial Literacy in Increasing Subscription to Shariah-Compliant Financial Products in Nigeria

In recent years, the demand for Shariah-compliant financial products has grown steadily in Nigeria, reflecting the increasing awareness of ethical and faith-based financial systems.However, despite this growth, the level of participation and subscription to Islamic financial products remains relatively low compared to the country’s Muslim population. This gap highlights a fundamental issue of limited Islamic financial literacy, which affects individuals’ ability to understand, trust, and adopt Islamic banking and finance solutions. Therefore, this study seeks to examine the role of Islamic financial literacy in increasing the subscription to Shariah-compliant financial products in Nigeria. Using a quantitative approach, primary data were collected through a structured questionnaire administered to 100 respondents across the states. Descriptive statistics assessed levels of Islamic financial literacy, while correlation analysis examined its relationship with product subscription. The findings indicate high awareness and basic understanding of Islamic finance principles, particularly the prohibition of interest (riba). However, the results reveal a weak and statistically insignificant relationship between Islamic financial literacy and subscription to Shariah-compliant financial products (r = –0.223, p = 0.150). This suggests that financial literacy alone does not drive adoption. Other factors, including accessibility, trust, religiosity, and perceived convenience, appear more influential. The study concludes that while Islamic financial literacy is important, it must be complemented by institutional and market-based strategies to enhance uptake in Nigeria.

An Ecolinguistics Analysis of Print Media Narratives on Climate Change and Disasters: A Case Study of the 2025 Floods in Pakistan

Climate change is intensifying extreme weather events globally, with countries like Pakistan facing severe vulnerability. Print media plays a critical role in constructing public understanding of these disasters. However, existing research on Pakistani media's disaster reporting primarily utilizes content analysis, overlooking the deep ideological work of language. A significant gap exists in applying a systematic ecolinguistics framework to analyze how linguistic constructions influence ecological awareness. This study aims to conduct an ecolinguistic analysis of Pakistani print media narratives on the 2025 floods to uncover the underlying "stories-we-live-by" concerning climate change and human-environment relationships. A qualitative research design will be employed, building a specialized corpus of articles from English and Urdu newspapers published during the crisis. Using Stibbe's (2015) ecolinguistic framework, the analysis will critically examine linguistic features, including metaphors, transitivity, and appraisal patterns, to identify dominant discourses. The analysis is expected to reveal a prevalence of anthropocentric and "destructive" discourses that frame nature as a threat and obscure systemic causes of climate change. Conversely, beneficial, ecocentric narratives promoting sustainability are anticipated to be marginal. The findings will offer crucial insights for media literacy, guiding journalists and policymakers towards more responsible, ecologically aligned climate communication that fosters public engagement and effective action in Pakistan.

Sentiment Analyses of Linguistic Ostracism and Employee Voice Behavior in Pakistani Employees Working Abroad.

Linguistic ostracism has emerged as a significant psychosocial challenge for immigrant employees, affecting their emotional well-being, social integration, and workplace effectiveness. Between 2022 and 2025, approximately 2.9 million Pakistanis migrated abroad for employment, making language-related exclusion a critical issue in expatriate work contexts. Despite growing attention to workplace ostracism, limited research has examined linguistic ostracism in relation to employee voice behavior among Pakistani expatriates. In particular, there is a notable lack of studies employing computational text analysis or sentiment-based approaches to capture the emotional dimensions of this phenomenon. This study addresses this gap by applying Natural Language Processing (NLP) and lexicon-based sentiment analysis to explore the emotional and behavioral dynamics of linguistic ostracism and employee voice behavior. Data were collected through an open-ended questionnaire from 70 Pakistani employees working abroad and analyzed using R-Studio. A cross-sectional research design was adopted, and aspect-based sentiment analysis was used to identify emotional patterns. Additionally, correlation and regression analyses were conducted to examine the relationship between linguistic ostracism and employee voice behavior. Findings indicate that linguistic ostracism is associated with predominantly negative emotional experiences among overseas Pakistani employees, which correspond with reduced employee voice behavior and increased prohibitive communication tendencies. However, statistical analysis revealed no significant direct relationship between linguistic ostracism and employee voice behavior, suggesting a more complex and indirect emotional–behavioral mechanism. Overall, the study contributes to design AI based language translating tools and emotionally intelligent virtual companions for foreign employees to promote inclusion and voice behavior in multicultural organisations

Sentiment Analysis of Public Reaction to the May 9, 2025 Tri-Service Joint Press Conference of Pakistan Armed Forces

This study analyzes public reactions to the May 2025 Tri-Service Joint Press Conference using sentiment analysis of social media comments. A mixed-method approach combines lexicon-based sentiment classification with thematic analysis to identify emotional patterns. Findings show predominantly positive and neutral responses, reflecting patriotism, trust, and national unity. The study also highlights the influence of linguistic framing and tone on public perception.

A Corpus Based Morphosemantic Analysis of Reduplication Patterns in Pahari

Reduplication is a productive morphosemantic phenomenon of most Indo-Aryan languages, which has significant grammatical and expressive roles. However, in Pahari, the same is yet to be documented, though this is commonly used in ordinary conversation. Current linguistic descriptions consist of a patchwork of descriptions or provide examples as opposed to systematic and empirical studies of reduplicative forms in this low-resource language. This paper fills this gap by carrying out corpus-based research on reduplication in Pahari. The study aims at establishing the most significant elimination of reduplications in the Pahari and analyzing their semantic senses and functions. Precisely, the study examines the role of reduplicative patterns in the creation of emphasis, intensity, iteration, and expressive nuance in natural language use. Corpus-based methodology is used (qualitative). An independent Pahari corpus composed of written and transcribed texts is examined with AntConc in order to find full, partial, and echo as well as ablaut reduplication. All of them will be analyzed by means of morphosemantic annotation and contextual interpretation in order to identify their communicative role. The research is likely to show that Pahari has a systematic system of reduplication, which will prove that reduplicative forms are important in defining meaning, boosting expressiveness, and compensating for inflectional morphology. The results will be useful in descriptive linguistics, as well as in documentation of a low-resource language of the Indo-Aryan branch, and future computational studies of the Pahari language.

Building a Specialized Corpus of Hausa Facebook Texts for Sentiment Analysis

Development of a specialized, normalized, and sentiment-annotated corpus of 2,000 Hausa Facebook posts/comments to bridge the low-resource language gap in NLP.

The Role of Corpus-Based Forensic Linguistic Profiling in Identifying Suspects in Cyber-Blackmail Communication

Cyber-blackmail is an increasingly prevalent digital crime in Pakistan, where offenders exploit anonymity to threaten, coerce, and manipulate victims for financial or personal gain. Traditional investigative methods often prove insufficient when perpetrators use anonymization tools, making linguistic evidence a crucial complementary resource. This study examines the linguistic characteristics of cyber-blackmail messages and evaluates the usefulness of corpus-based forensic linguistic profiling in identifying potential suspects. Using a dataset of 20 cyber-blackmail cases and applying Coulthard and Johnson’s (2007) four-stage framework Description, Comparison, Classification, and Identification; the analysis identifies recurrent lexical, syntactic, and pragmatic patterns across the corpus. Key findings reveal a shared repertoire of violent threats, reputational intimidation, abusive expressions, religious oaths, and coercive commands that shape a culturally embedded genre of cybercrime discourse. Comparative analysis demonstrates that while cyber-blackmail communication is highly formulaic, distinctive idiolectal markers such as repeated curse clusters, stable emotional exaggerations, and specific threat structures may assist in suspect narrowing. The study concludes that corpus-based forensic linguistic analysis provides systematic, replicable, and context-sensitive insights that enhance the interpretation and investigative value of cyber-blackmail communication. Future research should incorporate larger multilingual corpora and multimodal data to strengthen forensic applicability.

The Role of Corpus-Based Forensic Linguistic Profiling in Identifying Suspects in Cyber-Blackmail Communication.

Cross-Cultural Pattern of Female's Silence and Speech: A Comparative Corpus-Based Analysis of American and Pakistani Novels

A woman’s silence echoes louder than her words. The patterns of female silence and speech remain a debated issue in literature, whether it is due to oppression, resistance, or any form of trauma. Most previous studies rely on the qualitative interpretation of texts by examining them thoroughly to explore themes related to gender dynamics, male superiority and women's empowerment. However, limited attention is given to data-driven methods that can provide empirical data based on linguistic evidence. This research aims to do a comparative study of selected Pakistani and American novels by using corpus linguistics to explore recurring patterns of female silence and speech. A mixed-method approach will be employed in this research. A purposive corpus will be generated based on four selected novels. The empirical data will be collected by using AntConc software in the form of frequencies of male and female dialogues, collocations and repeated lexical patterns. This quantitative data will be further analyzed through the lens of Feminist Discourse Analysis to explore the contribution of patriarchal structure, power dynamics, and social and cultural constraints in affecting women’s ability to speak or to remain silent. This research concluded that social structure and cultural norms play a crucial role in suppressing or empowering the voice of women.

The Use of Deep Learning for Automatic Recognition of Urdu Dialects: A Speech Processing Study of Pakistani Speakers

Language variation plays an essential role in communication among Urdu speakers in Pakistan. Different dialects strongly affect mutual intelligibility and pose challenges for technological applications, including automatic speech recognition, voice-based interfaces, and language learning solutions. Understanding these variations is essential for developing accurate and inclusive language technologies that reflect Pakistan’s rich linguistic diversity. Most Urdu speech recognition research focuses on standard Urdu, neglecting regional dialectal variation, which limits accuracy and practical usability across regions. This study proposes a deep learning-based methodology for the automatic identification of Urdu dialects in natural speech and categorisation of regional differences in dialect-aware speech technologies. A quantitative experiment is conducted using a dataset of labelled audio recordings from native speakers across multiple regions. The dataset consists of 80 audio samples from different regions. Embeddings are extracted from the audio signals. To assess the performance of speech recognition, a convolutional neural network with bidirectional recurrent layers (CNN-BiRNN) is trained and assessed using performance measures, including accuracy, precision, recall and F1-score. The model has achieved 47% baseline accuracy on an imbalanced, low-resource dataset. This work contributes to computational linguistics, inclusive language technologies, and the creation of comprehensive Urdu language resources.

The Influence of System 2 Prompting in Urdu Generative Language Models: A Quantitative Trade-Off Study of LLM Fairness and Utility

Large language models (LLMs) these days exhibit impressive generative capabilities across many languages around the world, but still continue to reproduce gender stereotypes rooted in their training corpora and low-resource languages, such as Urdu, serve as an extremely hard obstacle for them for having morphological complexities. While structured “System 2” prompting has emerged as an accessible mitigation technique, its fairness-utility dynamics remain unquantified and unstudied in Urdu. In this study the effect of system 2 prompting is studied on the utility and bias in the responses of Qwen 7B, which is an open source LLM widely used for multilingual tasks. It involves testing of 100 prompts (utility=50, fairness=50) through two conditions: baseline, and system 2. Fairness was evaluated using a Gendered Term Association (GTA) metric, while the utility was assessed using the Factual Coherence and Linguistic Quality (FCLQ) scale. Results show a moderate baseline bias (with Fairness Score of 52%) and a 10-point fairness increase under System 2 prompting (an increase to 62%), with a marginal 0.22 decline in utility. A Trade-Off Ratio of 45.45 was found which indicates highly efficient fairness gain per unit utility cost. Upon further analyses, four distinct categories were observed in the responses: remained unbiased, mitigation success, stubborn bias, and harm/regression which demonstrated systematic variation in model behavior. This study provides a comprehensive, and quantitatively grounded fairness–utility benchmark for Urdu LLMs and provides practical and constructive insights for building ethically aligned and fair AI.

The Role of Language Framing in Communicating Financial Inclusion: A Discourse Analysis Study of Northern Nigeria

This paper analyses the application of language framing in the discourse on financial inclusion in Northern Nigeria, a multilingual society characterised by low digital literacy, religious sensitivities, and past marginalisation. Although current studies on financial inclusion in Nigeria have primarily focused on access, technology, and economic performance, the communicative mechanisms through which inclusion is facilitated and challenged have received insufficient attention. The paper draws on framing theory and Critical Discourse Analysis (CDA) to examine how linguistic choices shape people's cognitive, affective, and behavioral responses to digital financial services. Based on Fairclough's three-dimensional CDA framework, the study will examine media writings, institutional policy reports, and semi-structured interviews with policymakers, financial services providers, and community members conducted between 2012 and 2024. The analysis has shown that the concept of financial inclusion is largely defined by the language of empowerment, modernisation, and national development, using authoritative and technocratic terms in institutional discourses, and by the relational language of storytelling, communal pronouns, and Hausa-English code-switching in media discourses. Nevertheless, the community's responses show scepticism, including mistrust, fear of surveillance, and concerns about the compatibility of religion and cultural misalignment. The research concludes that language framing plays a significant role in shaping people and that culturally grounded, dignifying, and trust-based communication will be key to achieving inclusive financial participation in Northern Nigeria.

The Role of Language in Improving Plant Science Learning Among First-Year Undergraduate Botany Students in Sindh, Pakistan

This study investigated the linguistic challenges faced by first-year undergraduate Botany students in Sindh, Pakistan, assessed the impact of these challenges on learning and motivation, and documented the self-directed coping strategies students employ. Using a mixed-methods design with 45 students, results revealed that 77.8% acknowledged language barriers hindered comprehension, and 37.8% experienced reduced confidence. Students relied heavily on informal coping mechanisms (vernacular notes, online searches, peer translation) over formal instructor consultation (26.7%). The study concludes that integrating language-sensitive teaching practices is essential.

â€œThe Psycholinguistics Impact of Urdu-English Code Switching in Mobile Notification among Pakistani Youthâ€

In an increasingly digitized world, mobile notifications have become a primary gateway for delivering information across sectors such as banking, education, healthcare, and social media. The effectiveness of these notifications largely depends on their linguistic structure, timing, and clarity, which influence users’ comprehension and responsiveness. In a bilingual context like Pakistan, where Urdu and English are frequently used interchangeably, code-switching plays a significant communicative role. While existing research highlights the importance of code-switching in education, media, and identity formation, limited attention has been given to its impact within mobile notification systems—particularly in relation to psycholinguistic outcomes. investigating how different language formats—Urdu-only, English-only, and mixed Urdu-English—affect comprehension, perceived urgency, and compliance among Pakistani youth aged 18–35. Drawing on sociolinguistic and psycholinguistic perspectives, the research examines whether bilingual notifications enhance or hinder users’ ability to quickly understand and act upon information. Additionally, the study explores the moderating effects of demographic variables such as age and gender. The findings are expected to contribute to both theoretical and practical domains by offering insights into language use in digital communication and informing the design of more effective notification systems, particularly in critical areas such as health, emergency services, and financial security.

TITLE An Interpretive Phenomenological Exploration of Pakistani ESL Learnersâ€™ Disengagemental Dynamics in Contextual Critique

Contextually disengagement is the most common problematic area of ESL learners around the non-native world of the 21st century. It is a cause, which effects the learning proficiency, and performance within ESL classrooms. While previous researches had focused on the presence of the problematic situations in the learning as well as acquiring process of the ESL students. These studies also had suggested that autonomy, emotional support, and culturally responsive teaching could improve engagement but maximum research yet oversees the learners’ voice. This study aimed to focus on the learners’ autonomy, emotional support and cultural response through which they expressed their insights and emotional relationship with their second language. NVIVO was used for the thematic and discourse analysis. A qualitative methodology (Interpretive Phenomenology) was used to investigate ESL learners’ experiences in disengagement during contextual critique especially that dynamics which were related with emotional, cultural, and instructional factors contribute to learners’ hesitation and withdrawal in analytical classroom tasks. A distinctive feature of this study was the use of reflective metaphors, which served as expressive tools for participants to convey complex emotions and perceptions. It was expected that this study had identified the core disengagemental dynamics, focused on academic challenges, as well as offered grounded insights. Eventually, the study was expected to foster a reflective mindfulness among educators, policymakers, and curriculum designers about the complications, and complexities of ESL learners’ authentic needs as well as experiences

The Role of Interactional Meta discourse in Portraying Female Identity: A Corpus-Based Study of Fantasy TV Characters

This study explores the role of interactional metadiscourse in portraying female identity in Sci-Fi and fantasy television series. Using a corpus-based approach, the research analyzes the dialogue of five selected female characters to identify patterns of hedges, boosters, attitude markers, self-mention, and engagement markers based on Hyland’s (2005) model. The study aims to understand how these linguistic features reflect emotions, identity, and character development in fictional contexts.

The Role of Interactional Metadiscourse in Portraying Female Identity: A Corpus-Based Study of Fantasy TV Characters

The Impact of Artificial Intelligence on Undergraduate Studentsâ€™ Academic Performance in Pakistani Universities

Artificial Intelligence (AI) is transforming modern education by improving learning efficiency and academic performance. This study examines the role of AI among undergraduate students in Pakistani universities, highlighting both benefits and risks such as dependency and reduced critical thinking.

The Influence of Linguistic Determinants in Digital Self-Narratives on Suicidal Cognition: A Quantitative Study of Adolescents in Jhelum Valley

Adolescents’ digital self-narratives offer critical insight into their emotional and cognitive states, with particular relevance for detecting suicidal cognition in under-researched contexts like the Jhelum Valley. Existing studies often explore psychological or sociological factors but overlook measurable linguistic indicators in adolescents' self-narratives that can predict suicidal cognition. This study investigates the impact of linguistic determinants in adolescents’ digital self-narratives and examines their association with suicidal cognition, using a corpus-based, quantitative research design. A purposive sample of 120 publicly available adolescents’ digital self-narratives (P1–P120) was collected, resulting in a cleaned corpus of 1,489 words. Linguistic features were extracted using Linguistic Inquiry and Word Count (LIWC), evaluated with descriptive statistics (frequency and percentages), and interpreted using Beck's Cognitive Theory of Suicide and the Sapir-Whorf hypothesis. The findings indicate a strong self-referential profile and reveal patterns associated with negative emotional states, rumination, and absolutist thinking. The study supports the use of linguistic determinants in culturally sensitive early detection frameworks and digital well-being programs.

A Corpus-Based Analysis of Temporal Markers of Selected Afro-American Neo-Slave Narratives.

The study explores the use of temporal markers in Afro-American Neo-Slave narratives through a corpus-based approach. It examines how temporal markers reflect transition from trauma to healing in literary texts. Both quantitative and qualitative methods are used to identify the frequencies and to interpret the meanings of these temporal markers

Destination Corpus Research Center (CRC), Minhaj University Lahore

This study applies corpus-based and NLP-driven sentiment analysis to examine linguistic framing and ideological bias in global news coverage of Ukrainian migrants (2023–2025). Using sentence-level polarity scoring and lexical frequency analysis, the study compares Western and non-Western media to identify sentiment patterns and framing differences.

The Impact of Intergenerational Gap on Efik Language Use among Older and Younger Speakers: A Sociolinguistics Approach

This study investigates the impact of the intergenerational gap on Efik language use among older and younger speakers from a sociolinguistic approach. It examines the speech styles of both generations, identifies communication barriers arising from differences in language use, and proposes strategies for improving intergenerational communication and preserving the Efik language. The study is guided by Dell Hymes’ Ethnographic Model of Speaking and adopts a qualitative research approach. Data were collected through purposive sampling from Efik speakers in Calabar South and Calabar Municipality using interviews, questionnaires, and participant observation. The findings reveal clear differences in fluency, vocabulary, and communicative styles between generations. Older speakers demonstrate greater fluency and frequently use traditional expressions and proverbs, while younger speakers often employ slang and code-mixing influenced by English and Nigerian Pidgin. These differences contribute to communication barriers and reduced transmission of cultural knowledge. The study emphasizes the importance of strengthening cultural education and intergenerational interaction to support the preservation and revitalization of the Efik language.

Evaluating Forced Alignment in an Under-Documented Shakargarh Punjabi Speech Corpus

This project evaluates forced alignment in an under-documented Shakargarh Punjabi speech corpus. It aims to test how effectively forced alignment performs on naturally occurring speech from a local Punjabi-speaking community, identify the main factors affecting alignment accuracy, and assess the usefulness of aligned speech data for corpus development and language documentation.

A Corpus-Based Assessment of Accuracy, Readability, and Lexical Complexity in AI-Generated versus Human-Authored Drug Summaries

With the increasing popularity of artificial intelligence (AI), people have started to use it in many areas of their lives. Many users have become dependent on it for their daily tasks and even use it for medical advice. People now use AI to get information about medicines, including their uses, indications, side effects, and doses. As more people begin to rely on AI for such information, it is important to study whether the information provided by AI is accurate and easy to read. Only a few studies have examined this, and most of them focused only on language-based variables like readability or used very small samples. This study aimed to analyze the accuracy, readability, and lexical complexity of AI-generated and human-authored drug summaries to see which one was more accurate and easier to read. A mixed-method approach was used. A corpus of 90 drug summaries was created by selecting 30 drugs from 10 drug classes. For each drug, three summaries were collected from ChatGPT, Gemini, nd trusted human sources such as Drugs.com and Medscape.com. The corpus was analyzed using different formulas and tools to check readability (Flesch Reading Ease, Kincaid, SMOG, Fog, ARI) and lexical complexity (Lexical Density, Lexical Diversity, Lexical Sophistication). Accuracy was checked manually. The results of this study show that human-authored drug summaries are more accurate (100%) as well as more readable, with higher FRE scores and lower grade-level scores. About 12–14 years of education is required to understand the responses from human sources. On the other hand, AI performed less effectively than humans. Gemini performed slightly better than ChatGPT. The accuracy of AI ranged between 71%–76%. Other variables showed that AI-generated summaries are more difficult and complex to understand. At least 14 17 years of education is required to understand Gemini responses, while 15–18 years of education is required to understand ChatGPT responses.

Recent Posts

Where Vision Meets Concrete Reality

Raising the Bar in Construction.

Get In Touch

Research Project