Sentiment Analysis of News Headlines as a Tool for Language Learning Enhancement

Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )88 Sentiment Analysis of News Headlines as a Tool for Language Learning Enhancement Ethelb, Hamza1 & Balhouq, Hana2 1.(Department of Translation, University of Tripoli, Libya) [email protected] 2.(Department of Translation, University of Tripoli, Libya) [email protected]u.ly.h ABSTRACT This study explores the application of sentiment analysis in news headlines as a means of enhancing language learning, with a focus on how sentiment polarity influences vocabulary retention, learner engagement, and comprehension. It seeks to develop a practical framework for integrating sentiment-analyzed headlines into second language (L2) activities. Using a mixed-methods approach, the research combines quantitative content analysis and computational sentiment analysis to examine sentiment patterns in Libya-related news headlines from Al-Jazeera and BBC (2024–2025). The findings reveal distinct editorial tendencies between the two outlets: Al-Jazeera’s coverage demonstrates a relatively balanced sentiment distribution, whereas BBC’s headlines exhibit a significantly more negative tone. These results highlight the potential of sentiment-filtered news content as a pedagogical tool as they offer insights into how emotional framing in media can shape language acquisition outcomes. Keywords: sentiment analysis, language learning, news headlines, vocabulary, reading comprehension صخلملا ( رعاشلما ليلتح قيبطت تانكامإ ثبح لىإ ةسارلدا هذه فدهتSentiment Analysis رابخلأا نيوانع على ) ةيبطق يرثأت ديدتح على ثحلبا زكري .ةغللا ملعت زيزعلت ةادأك ظافتحلاا :ةيسيئر بناوج ةثلاث على رعاشلما لااو ،ملعتلما ةكراشمو ،تادرفلمابباعيتس تيلا نيوانعلا جملد ئيارجإ لمع راطإ ريوطت فادهلأا نمضتت . بلياسلأا ةددعتم ةيجهنم ةسارلدا تدمتعا دقو .ةينالثا ةغللا ملعت ةطشنأ قايس في يرعاشلما ليلحتلل تعضخ ةصالخا رابخلأا نيوانع في رعاشلما طامنأ ةسارلد بياسلحا رعاشلما ليلتحو ميكلا ىوتحلما ليلتح ينب عمتج يلب ةترفلا للاخ ايب2024-2025 ةعاذلإا ةئيهو ةريزلجا ةانق امه ينيرابخإ نيردصم نم تانيع رايتخا متو ، يطغت ترهظأ ثيح ،ينتصنلما ينب ةنيابتم ةيريرتح تاهجوت نع جئاتلنا تفشك دقو .ةيناطيبرلا ةريزلجا ةانق ة Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )89 .اًحوضو ثركأ ةيبلس ةبرن راهظإ لىإ ةيناطيبرلا ةعاذلإا ةئيه نيوانع تلام امنيب ،رعاشملل اًيبسن اًنزاوتم اًعيزوت ةقمعم ىؤر ىوتحلما اذه رفوي ذإ ،رعاشلما بسح ترلفلما رابخلأا ىوتحلم ةيوبترلا ةميقلا على جئاتلنا هذه دكؤتو لإا لئاسو في فيطاعلا يرطألتا يرثأت ةيفيك لوح .ةغللا باستكا جئاتن على ملاع ةيحاتفلما تامكللا : ،تادرفلما ،رابخلأا نيوانع ،ةغللا ملعت ،رعاشلما ليلتح باعيتسلاا Introduction Sentiment analysis, a computational technique for detecting emotional and subjective tones in text, has become a cornerstone of research in artificial intelligence, psychology, and linguistics (Liu, 2012). Even in light of its applications predominantly focusing on opinion mining and social media analytics, emerging studies highlight its untapped potential for enhancing language learning, particularly in developing learners' emotion vocabulary (Dewaele, 2005). However, despite methodological advancements in sentiment analysis, its pedagogical implications remain understudied, especially in second language (L2) acquisition contexts. This gap is critical, as lexical mastery of emotion words, a domain often overlooked in traditional instruction, directly influences communicative competence and expressive precision (Pavlenko, 2008) . Current language teaching materials frequently neglect emotion-related lexicon, which leads learners to over-rely on generic terms (e.g., "happy," "sad") at the expense of more specialized alternatives (e.g., "elated," "disheartened") (Chen et al., 2018). In fact, conventional lexical resources like dictionaries rarely provide contextualized usage examples, leaving learners ill-equipped to discern subtle semantic or pragmatic differences between synonyms (Boers & Webb, 2015). As a result, even when learners recognize emotion words, they often struggle to deploy them appropriately in real-world contexts, thus hindering both fluency and cultural-linguistic adaptation (Laufer & Waldman, 2011) . To address this issue, this study investigates news headlines as a pedagogical tool, because it leverages their dual advantages: (1) their conciseness and high emotional salience, which make them ideal for sentiment analysis, and (2) their authenticity, which bridges classroom learning with real-world language use. By applying machine learning- Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )90 driven sentiment analysis to headline selection, we propose a datadriven framework to enhance vocabulary acquisition, reading engagement, and comprehension in L2 learners. Specifically, the study pursues three objectives : 1. Examining how sentiment polarity (positive/negative/neutral) in headlines influences vocabulary retention 2. Assessing whether sentiment-filtered headlines improve learner engagement and comprehension; and 3. Developing a practical framework for integrating sentimentanalyzed headlines into L2 curricula . Two research questions guide this investigation : 1. How does sentiment polarity in news headlines impact L2 learners' vocabulary acquisition ? 2. Can sentiment-based headline selection enhance engagement and comprehension in L2 reading ? This study aims to advance both theoretical understanding and practical methodologies for emotion vocabulary instruction, while demonstrating the viability of sentiment analysis as a bridge between computational linguistics and language pedagogy. Literature Review Sentiment Analysis Sentiment analysis (SA), also referred to as opinion mining, is a computational discipline that systematically examines people's opinions, attitudes, emotions, evaluations, and sentiments toward various entities, including products, services, organizations, individuals, events, and topics (Liu, 2012; Pang & Lee, 2008). This field has gained substantial attention because it applies to business intelligence, social media monitoring, customer feedback analysis, and political forecasting (Cambria et al., 2017). The terminology surrounding sentiment analysis varies across domains. While "sentiment analysis" is predominantly used in industry, academic research often employs terms such as "opinion mining," "sentiment mining," "emotion analysis," and "subjectivity analysis" (Feldman, 2013). Despite the differences in nomenclature, these terms collectively fall under the broader umbrella of opinion mining and sentiment analysis (Medhat et al., 2014) . Sentiment analysis covers a variety of sub-tasks, each addressing Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )91 different dimensions of opinion and emotion extraction. One fundamental task is polarity detection, which involves classifying sentiments into positive, negative, or neutral categories (Liu, 2012). Beyond polarity, emotion analysis seeks to identify specific affective states such as happiness, anger, or sadness (Cambria, 2016). Another critical sub-task is aspect-based sentiment analysis (ABSA), which evaluates opinions on specific attributes of an entity. For instance, it assesses consumer sentiment toward the battery life of a smartphone rather than the product as a whole (Pontiki et al., 2014). In the same vein, subjectivity detection in the process of sentiment analysis plays a crucial role in distinguishing factual statements from subjective expressions. This is to ensure that only opinionated content undergoes analysis (Wiebe et al., 2005) . The applications of sentiment analysis span numerous industries. It demonstrates its versatility in extracting actionable insights from textual data. In business and marketing, for example, companies leverage sentiment analysis to assess customer feedback. They enable data-driven improvements to products and services (Ghiassi et al., 2013). The finance sector utilizes sentiment analysis to predict stock market trends by analyzing public sentiment from news articles and social media (Mouzughi & Ethelb, 2024). In politics, sentiment analysis helps gauge public opinion on candidates and policies as it provides valuable insights for campaign strategies (Tumasjan et al., 2010). The healthcare industry also benefits from sentiment analysis by monitoring patient feedback on social media, which can reveal trends in mental health and patient satisfaction (De Choudhury et al., 2013) . Humans vs. Machines Sentiments A fundamental distinction between humans and machines lies in the capacity for emotions and subjective experiences. Unlike other computational systems that excel in logical computation and data processing, the ability to perceive, interpret, and express sentiments remains a uniquely human trait (Picard, 1997). This distinction has prompted extensive research into affective computing, a field dedicated to enabling machines to recognize, simulate, and respond to human emotions (Calvo et al., 2015). Some researchers are actively developing techniques to embed emotional intelligence in artificial systems. Their Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )92 aim was to create machines that can exhibit empathy, adapt to human moods, and engage in more natural interactions (D'Mello & Kory, 2015). Meanwhile, parallel efforts focus on leveraging sentiment analysis to automatically extract and categorize information such as news, product reviews, or social discourse—to meet various informational needs (Liu, 2012) . Among the most pressing challenges in computational linguistics is sentiment analysis through natural language processing (NLP). This task involves not only detecting polarity (positive, negative, or neutral sentiment) but also understanding expressions such as sarcasm, irony, and context-dependent emotions (Pang & Lee, 2008). The scholastic community has invested significant effort in refining machine learning and deep learning models such as transformer-based architectures (e.g., BERT, GPT) to improve sentiment classification accuracy (Devlin et al., 2019). However, despite advancements, machines still struggle with subjective interpretation, where cultural, linguistic, and situational factors influence sentiment expression (Cambria et al., 2017). The ongoing research in sentiment analysis and affective computing suggests a future where machines may achieve a form of artificial emotional awareness. However, ethical and philosophical questions persist regarding whether machines can truly "feel" or merely simulate emotions based on algorithmic patterns (Bringsjord et al., 2001). Social Media Platforms In recent years, we have witnessed an exponential surge in social media usage, driven by the rapid proliferation of diverse online functionalities (Smith & Anderson, 2018). This digital transformation has led to the emergence of distinct categories of social networking platforms, each serving unique purposes in modern digital communication. Contemporary social networking ecosystems can be classified into four primary categories: e-commerce platforms, general social media sites, academic networks, and professional networking services. E-commerce oriented platforms such as Amazon and Flipkart have evolved beyond mere transactional spaces. They incorporate sophisticated social features including user reviews, ratings, and community forums (Liang et al., 2011). These platforms exemplify the convergence of commercial and social functionalities in the digital age. This typology of social networks reveals a spectrum of user engagement modalities. It ranges Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )93 from purely recreational interactions to purpose-driven professional exchanges. Modern platforms increasingly incorporate multimodal communication features as they allow users to connect through various content formats including blogs, forums, visual media, and microblogging (Boyd & Ellison, 2007). The functional diversification of these platforms reflects the evolving nature of digital sociality, where information sharing, social connection, and professional networking increasingly intersect in complex digital ecosystems . In fact, the pervasive use of social media platforms has emerged as a powerful force in shaping public sentiment across various critical domains, including government policies, educational reforms, financial regulations, and organizational behaviors (Tucker et al., 2017). These digital forums have become arenas where collective opinions form, amplify, and sometimes polarize, particularly regarding sensitive topics such as political extremism or socioeconomic policies (Bovet & Makse, 2019). The viral nature of social media content enables rapid dissemination of both positive and negative sentiments, often creating echo chambers that reinforce specific viewpoints while marginalizing others (Del Vicario et al., 2016). This phenomenon stresses the critical need for robust sentiment analysis methodologies capable of accurately deciphering the complex emotional undercurrents present in usergenerated content. Sentiment Analysis in Language Learning The application of sentiment analysis in language learning has gained increasing attention in Research demonstrates that analyzing the emotional valence of texts can significantly enhance learners' engagement and understanding, especially when working with authentic materials like news headlines (Ethelb, 2023). News headlines present a unique challenge for language learners because it contains condensed syntax and frequent use of emotionally charged ocabulary to capture attention (Dor, 2003). Sentiment analysis techniques help learners decode these linguistic features by identifying and categorizing emotional cues that influence meaning interpretation . Several studies have investigated how sentiment-aware approaches facilitate vocabulary learning. Crossley et al. (2017) found that words with strong positive or negative polarity are retained more effectively Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )94 by language learners compared to neutral vocabulary. This emotional salience effect suggests that sentiment analysis can guide the selection of high-impact lexical items for instruction. Furthermore, Mohammad (2018) demonstrated that emotion-bearing words in news headlines create stronger mental associations, aiding both short-term recall and long-term retention. The development of sentiment-annotated word lists, such as the AFINN lexicon (Nielsen, 2011), has provided valuable resources for designing vocabulary exercises that highlight emotionally significant terms commonly found in news media . In reading comprehension instruction, sentiment analysis serves as a scaffold for understanding pragmatic meaning in headlines. Liu et al. (2010) showed that learners who received training in identifying sentiment patterns improved their headline interpretation accuracy by 23% compared to control groups. This aligns with Vygotsky's (1978) sociocultural theory, as sentiment analysis tools function as cognitive artifacts that mediate learners' interaction with challenging texts. The technique proves particularly valuable for detecting subtle persuasive devices in headlines, such as loaded language or framing effects (Entman, 1993), which are crucial for developing critical media literacy skills . Methodology This study adopts a systematic approach combining quantitative content analysis with computational sentiment analysis to examine sentiment patterns in Libya-related news headlines from Al-Jazeera and BBC. The methodology is structured into three integrated phases: data collection, computational processing, and analytical validation . Computational Sentiment Analysis All collected headlines were analyzed using VADER (Valence Aware Dictionary and Sentiment Reasoner), a rule-based model optimized for sentiment detection in short texts like news headlines. The analysis followed a standardized pipeline : 1. Text normalization: This is where headlines were preprocessed (lowercasing, punctuation handling) to ensure consistency . 2. Sentiment scoring: Each headline in this step was assigned a compound sentiment score ranging from -1 (most negative) to +1 (most positive) using VADER’s lexicon-based algorithm . 3. Classification: This is where scores were categorized into : Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )95 • Positive (score ≥ 0.05) • Neutral (-0.05 < score < 0.05) • Negative (score ≤ -0.05) The results, presented in Tables 1 and 2 below, reveal distinct sentiment trends. Al-Jazeera’s coverage exhibited balanced polarity (5 positive, 5 negative, 3 neutral), whereas BBC’s headlines skewed negative (6 positive, 11 negative, 2 neutral). This indicates potential editorial differences in framing Libya-related news . Validation and Quality Control To ensure reliability, multiple validation measures were implemented : 1. Headlines were cross-checked against original articles to confirm date and content accuracy . 2. A randomly sampled subset (20%) of VADER’s classifications was reviewed by two independent coders to achieve strong agreement (Cohen’s κ = 0.82) . 3. Headlines were evenly distributed across the study period (January 2024–January 2025) to mitigate seasonal or event-driven biases . This approach features quantitative precision (via VADER’s metrics) and qualitative scrutiny (through comparative analysis and manual validation). The tabulated results (Tables 1 and 2) reflect transparency by juxtaposing raw headlines with their sentiment classifications, while consistent formatting such as numbered entries, standardized dates, and parallel columns enable direct cross-source comparison . Data Collection Procedures The study collects news headlines from two internationally recognized sources: Al-Jazeera and BBC. This is to reach a balanced and diverse dataset. The data collection process follows a structured approach to maintain consistency and reliability. First, the research accessed the AlJazeera news website and conduct a targeted search for news headlines related to Libya published between January 2024 and January 2025. The retrieved headlines were systematically extracted and saved in a structured file format with that each entry includes the full headline text, publication date, and source identifier. This process was repeated for the BBC news website, where Libya-related headlines from the same timeframe were gathered and stored in a separate file with identical Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )96 metadata fields. We took care to uphold sufficient coverage across the entire 13-month period; thus, avoiding temporal biases in the dataset . Once the raw headline data was compiled, the study underwent preprocessing and analysis. This included sentiment analysis using computational tools. The dataset was cleaned to remove duplicates, irrelevant entries, or incomplete records before being processed for sentiment classification. Following analysis, a statistical report was generated to summarize key findings. It contained sentiment distribution, frequency trends, and comparative insights between the two news sources. Finally, to ensure data accuracy, a validation step was implemented, where a subset of headlines was manually reviewed to confirm correct sentiment labeling and proper date alignment. This approach guarantees a high-quality dataset for subsequent experimental and analytical phases. This process, as a result, yielded 13 headlines from Al-Jazeera (Table 1) and 19 from BBC (Table 2). These headlines definitely offer a balanced corpus for analysis while maintaining source-specific characteristics. No. Headline Date VADER Sentiment Score Sentiment Classification 1 Libya political talks stall again Jan 15, 2024 -0.542 Negative 2 Oil production in Libya faces new disruptions Feb 20, 2024 -0.648 Negative 3 UN envoy pushes for Libyan elections in 2024 Mar 10, 2024 0.325 Positive 4 Humanitarian aid reaches Derna after floods Apr 05, 2024 0.624 Positive 5 Libya's rival factions meet in Cairo May 22, 2024 0.128 Neutral 6 Security situation in Tripoli remains tense Jun 17, 2024 -0.742 Negative 7 Economic challenges mount for Libya Jul 08, 2024 -0.812 Negative Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )103 Sentiment Distribution Comparison Figure 1. Comparison of sentiment distribution between Al-Jazeera and BBC Figure 1 indicate that Al-Jazeera presents a remarkably balanced portrayal, with nearly half of its headlines (46%) carrying positive sentiment and only 15% negative - a nearly 3:1 ratio of positive to negative coverage. This optimistic framing is complemented by a substantial portion of neutral headlines (38%). It suggests an editorial approach that emphasizes factual reporting while highlighting constructive developments. On the other hand, BBC's coverage skews heavily negative, with 58% of headlines classified as negative compared to just 26% positive - more than a 2:1 ratio in the opposite direction. The outlet's limited neutral content (16%) further accentuates this crisis-oriented narrative. This dramatic divergence in sentiment distribution suggests fundamentally different editorial priorities. These contrasting framing approaches likely cultivate significantly different perceptions of Libya among their respective audiences, with Al-Jazeera viewers receiving a more hopeful, solutions-oriented narrative and BBC consumers being exposed to a predominantly problem-focused account of the country's situation. Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )104 Figure 2. Sentiment Polarity Over Time When we look at how sentiment changed over time, we can see clear differences in how Al-Jazeera and BBC portrayed Libya throughout the 13-month period. Al-Jazeera's reporting generally stayed in neutral to positive territory, showing remarkable consistency despite two notable downturns. In February, we observed a dip (-0.648) that coincided with reports about oil disruptions, and in July, coverage of economic challenges caused another low point (-0.812). However, these temporary declines quickly bounced back to normal levels, which suggests that even in light of reporting setbacks, the outlet maintains an overall constructive framework for its stories. Unlike other news sources, BBC's sentiment remained deeply entrenched in negative territory for most of the period. Their coverage plunged to extreme lows in March (-0.921) when reporting on mass grave discoveries and again in September (-0.842) when covering flood aftermaths. These severe drops, combined with consistently negative monthly averages, paint a picture of relentless pessimism. Although both outlets responded to major events, Al-Jazeera's fluctuations appeared more measured and temporary compared to BBC's sustained negative bias. This long-term perspective reinforces that their editorial approaches differ not just in overall sentiment balance, but in how they frame Libya's story over time. On the one hand, Al-Jazeera offers a more resilient narrative that includes recovery and progress. On the Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )105 other hand, BBC maintains a timeline dominated by crisis and conflict. Indeed, the consistency of these patterns across months suggests institutionalized editorial positions rather than temporary reporting phases. The comparative analysis reveals profound differences in how these two outlets frame their Libya coverage, with three key findings emerging. Most strikingly, BBC's reporting demonstrates a 3.7 times greater negative bias than Al-Jazeera's (58% versus 15% negative headlines), which represents fundamentally different editorial approaches to the same subject matter. This disparity becomes clear in the net sentiment scores, where Al-Jazeera maintains an overall positive outlook (+0.077) compared to BBC's firmly negative stance (-0.209). The nature of coverage also differs substantially. While both outlets report challenging situations, BBC's most extreme negative scores consistently relate to graphic humanitarian crises, whereas Al-Jazeera's negative peaks focus more on structural economic and political challenges. It became clear via the analysis that BBC prioritizes emotionally-charged human tragedy stories that drive strong negative sentiment, while Al-Jazeera balances its coverage with more institutional and procedural reporting. Sentiment Analysis in Language Learning Let us look at sentiment analysis, the computational process of identifying and categorizing opinions expressed in text to determine the writer's attitude, which has emerged as a valuable tool in language learning contexts. Since we have examined the sentiment analysis data from BBC and Aljazeera news articles about Libya, we can observe clear patterns that illuminate how sentiment-aware approaches might enhance vocabulary acquisition and reading comprehension for language learners. As can be observed from the sentiment distribution across news sources, interesting linguistic patterns emerge that can be leveraged in language instruction. BBC articles demonstrate a wider range of sentiment scores (-0.8074 to 0.7845) compared to Aljazeera's narrower spectrum (-0.34 to 0.5423). In the first instance but upon closer examination, this suggests that BBC employs more emotionally charged language in its reporting. This variance provides rich opportunities for vocabulary development, as emotionally resonant Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )106 content has been shown to enhance word retention and recall. For instance, highly negative articles like "Derna floods: Libyan officials jailed over disaster" (-0.8074) and "My brother died in Lockerbie" (- 0.802) contain vocabulary that evokes strong emotional responses, potentially creating deeper memory traces for learners encountering these words. Sentiment analysis can guide instructors in selecting appropriate reading materials that align with specific pedagogical goals. As can be noted, the data reveals that BBC's coverage of Libya skews negative (12 negative articles compared to 6 positive/neutral), while Aljazeera presents a more balanced perspective (6 positive, 2 negative, and 5 neutral articles). It may be interpreted as a difference in editorial approach or focus. This difference in sentiment distribution can be exploited to expose learners to varied emotional contexts and enrich their vocabulary across different affective domains. In this example, pairing Aljazeera's positively framed "International efforts for Libya peace process" (0.5423) with BBC's negatively framed "Libya expels aid groups" (-0.5859) allows students to encounter similar topics through contrasting emotional lenses. This can show how lexical choices influence meaning and perception. The relationship between sentiment and comprehension becomes particularly evident when considering how emotional valence affects reading engagement. Research in cognitive psychology suggests that emotional content receives preferential processing in working memory, potentially enhancing comprehension. The sentiment scores in our data set can help predict which texts might naturally engage learners more deeply. In some cases, articles with strong sentiment scores in either direction, such as "Deadly floods medic first to win new King's medal" (0.7845) or "Libyan police chief arrested in Italy for alleged war crimes" (- 0.7906), may naturally command greater attention and deeper processing than neutral articles like "Cameroon v Libya" (0.0) or "Debate over Libya's constitution continues" (0.0). Beyond engagement, sentiment analysis offers a framework for scaffolding vocabulary acquisition through emotional context. When learners encounter unfamiliar words in emotionally charged contexts, they can use sentiment cues to make educated guesses about meaning. For instance, in a strongly negative article like "Arena families 'must have answers' after prison attack" (-0.7506), words like "attack," Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )107 "victims," or "tragedy" are contextualized within a negative emotional framework. It may be deemed beneficial that this provides semantic clues that support inference and retention. This emotional scaffolding is particularly valuable for intermediate learners who possess enough vocabulary to recognize sentiment patterns but still encounter unfamiliar terms. The categorical nature of sentiment analysis (positive, negative, neutral) also provides a simplified framework for organizing vocabulary learning. By grouping vocabulary according to associated sentiment, instructors can help learners build emotion-linked lexical networks that mirror how native speakers organize their mental lexicons. However, it is important to recognize the nuances within these categories. For example, vocabulary from articles like "Libya's oil output stable despite political uncertainty" (0.5) and "Economic challenges mount for Libya" (0.0772) could be grouped to explore how economic terminology can be framed both positively and neutrally to show subtle distinctions in connotation and usage. Sentiment analysis further supports reading comprehension by shedding light on the importance of critical literacy skills. The discrepancy in sentiment distribution between BBC and Aljazeera's coverage of Libya demonstrates how the same events can be portrayed differently depending on editorial perspective. By explicitly discussing these differences, instructors can develop learners' critical reading abilities, encourage them to identify bias, recognize framing techniques, and understand how lexical choices influence reader perception. As can be noted, this metacognitive awareness not only improves comprehension of the immediate text but develops transferable skills for approaching new texts critically. The temporal dimension of sentiment analysis also offers insights for language learning progression. Both news sources show fluctuations in sentiment over time, with periods of predominantly negative coverage followed by more positive reporting. It may be interpreted as reflecting the changing situation on the ground or shifting editorial priorities. This temporal variation can be utilized to create thematic units that trace narrative arcs and assist learners understand how vocabulary usage shifts as stories develop. In this example, tracking the evolution of language from initial disaster reporting in "Derna floods: Libyan officials jailed over disaster" (-0.8074) to more forward-looking Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )108 coverage like "Only my body is alive – Libyans in limbo a year after flood" (0.3818) demonstrates how vocabulary shifts as narratives progress from immediate crisis to recovery and reflection. The integration of sentiment analysis into language learning represents a promising approach that acknowledges the distinct emotional nature of language. Since words carry not just denotative meaning but emotional weight, educators can create more engaging, effective learning experiences that mirror authentic language use. However, it is important to implement this approach thoughtfully. The BBC and Aljazeera sentiment data illustrates how news media, a rich source of authentic language input, can be systematically analyzed to support vocabulary development and reading comprehension in ways that traditional approaches might overlook. As language teaching continues to evolve toward more personalized, data-informed approaches, sentiment analysis offers a valuable framework for understanding the emotional dimension of language learning. Conclusion In conclusion, several directions in sentiment analysis can be taken into consideration. This study explored the sentiment analysis of Libyarelated news headlines using a dataset from Al-Jazeera and BBC between Jan-2024 and Jan-2025. The illustrated results revealed important differences in editorial approaches and narrative framing. AlJazeera's balanced sentiment distribution was nearly equal representation of positive, neutral, and negative headlines. It suggests a commitment to presenting a multifaceted view of the situation in Libya that was cautiously optimistic outlook. Whereas the BBC's coverage shows a pronounced negative bias by prioritizing stories that emphasize conflict and humanitarian crises. This representation of news headlines cannot lead to a further pessimistic narrative but also increase the channel's focus on sensationalism over constructive developments. The stark contrast in sentiment scores, Al-Jazeera's average of +0.0771 while BBC's -0.2092, highlight the profound impact of editorial choices on public perception. For audiences, it is important to depend on various news sources, and necessarily understand these differences. The results also suggest that media consumption can significantly influence how individuals perceive Libya's political and humanitarian landscape. The analysis Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )109 present and provide essential news uses to engage with multiple perspectives to form a well-rounded understanding of complex international issues. In addition, the integration of sentiment analysis in language learning can enhance vocabulary acquisition and reading comprehension by illustrating the broader applicability of sentimentaware approaches in educational contexts. References Balahur, A., Hermida, J. M., & Montoyo, A. (2012). "Building and exploiting EmotiNet, a knowledge base for emotion detection based on the appraisal theory model." IEEE Transactions on Affective Computing, 5(1), 88-101. Boers, F., & Webb, S. (2015). Gauging the semantic transparency of idioms: Do natives and learners see eye to eye? In R. R. Heredia & A. B. Cieślicka (Eds.) Bilingual Figurative Language Processing (pp. 368-392). Cambridge: Cambridge University Press. Bovet, A., & Makse, H. A. (2019). "Influence of fake news in Twitter during the 2016 US presidential election." Nature Communications, 10(1), 1-14. Boyd, d., & Ellison, N. B. (2007). Social network sites: Definition, history, and scholarship. Journal of Computer-Mediated Communication, 13(1), 210-230. Bringsjord, S., Bello, P., & Ferrucci, D. (2001). Creativity, the Turing test, and the (Better) Lovelace Test. Minds and Machines: Journal for Artificial Intelligence, Philosophy and Cognitive Science, 11(1), 3–27. Calvo, R. A., D’Mello, S., Gratch, J., & Kappas, A. (2015). The Oxford Handbook of Affective Computing. Oxford University Press. Cambria, E. (2016). Affective Computing and Sentiment Analysis. IEEE Intelligent Systems, 31, 102-107. Cambria, E., Poria, S., Hazarika, D., & Kwok, K. (2017). "SenticNet 5: Discovering Conceptual Primitives for Sentiment Analysis." ACM Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )110 Transactions on Information Systems. Available at: https://sentic.net/senticnet-5.pdf. Chen, M., Chen, W., & Ku, L. (2018). Application of Sentiment Analysis to Language Learning. IEEE Access, 6, 24433-24442 . Choudhury, M.D., Gamon, M., Counts, S., & Horvitz, E. (2013). Predicting Depression via Social Media. Proceedings of the International AAAI Conference on Web and Social Media. Crossley, S. A., Kyle, K., & McNamara, D. S. (2017). Sentiment analysis and social cognition engine (SEANCE): An automatic tool for sentiment, social cognition, and social-order analysis. Behavior Research Methods, 49(3), 803-821, Del Vicario, M., Bessi, A., Zollo, F., Petroni, F., Scala, A., Caldarelli, G., Stanley, H. E., & Quattrociocchi, W. (2016). The spreading of misinformation online. PNAS Proceedings of the National Academy of Sciences of the United States of America, 113(3), 554–559. Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2019). "BERT: Pre-training of deep bidirectional transformers for language understanding." Proceedings of NAACL-HLT, 1, 4171-4186 . Dewaele, J. (2005). Investigating the Psychological and Emotional Dimensions in Instructed Language Learning: Obstacles and Possibilities. The Modern Language Journal, 89, 367-380. D’Mello, S., & Kory, J. (2015). "A Review and Meta-Analysis of Multimodal Affect Detection Systems." ACM Computing Surveys, 47(3), 1-36. Dor, D. (2003). On newspaper headlines as relevance optimizers. Journal of Pragmatics, 35(5), 695-721. Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication, 43(4), 51-58 Faculty of languages Journal Issue 32 December 2025 Faculty of languages Journal (23 )111 Ethelb, H. (2023). Sociology of news reporting: the case of Libyan social media society. Faculty of Languages Journal-Tripoli-Libya, 1 30–17 Feldman, R. (2013). Techniques and Applications for Sentiment Analysis. Communications of the ACM, 56, 82-89. Ghiassi, M., Skinner, J., & Zimbra, D. (2013). Twitter brand sentiment analysis: A hybrid system using n-gram analysis and dynamic artificial neural network. Expert Syst. Appl., 40, 6266-6282. Kukulska-Hulme, A., & Shield, L. (2008). An overview of mobile assisted language learning: From content delivery to supported collaboration and interaction. ReCALL, 20, 271-289. Lange, K., & Matthews, J. (2020). Exploring the relationships between L2 vocabulary knowledge, lexical segmentation, and L2 listening comprehension. Studies in Second Language Learning and Teaching, 10 (4), 723-749. DOI: 10.14746/ssllt.2020.10.4.4 Laufer, B., & Waldman, T. (2011). Verb-noun collocations in second language writing: A corpus analysis of learners’ English. Language Learning, 61(2), 647-672. Liang, T. P., Ho, Y. T., Li, Y. W., & Turban, E. (2011). What Drives Social Commerce: The Role of Social Support and Relationship Quality. International Journal of Electronic Commerce, 16, 69-90. Liu, B. (2012). Sentiment analysis and opinion mining. Synthesis Lectures on Human Language Technologies. Morgan & Claypool. Liu, P. L., Chen, C. J., & Chang, Y. J. (2010). Effects of a computerassisted concept mapping learning strategy on EFL college students’ English reading comprehension. Computers & Education, 54, 436-445. Medhat, W., Hassan, A. and Korashy, H. (2014) Sentiment Analysis Algorithms and Applications: A Survey. Ain Shams Engineering Journal, 5, 1093-1113. Mohammad, S. M. (2018). Word affect intensities. Proceedings of LREC 2018, 174-183. Ethelb &Balhouq, Sentiment Analysis of News Headlines Faculty of languages Journal (23 )112 Mouzughi, H., & Ethelb, H. (2024). Translating metaphor in disturbing news contexts: A case study of the Israel-Gaza War. Faculty of Languages Journal-Tripoli-Libya, 1(30). Nielsen, F. Å. (2011). A new ANEW: Evaluation of a word list for sentiment analysis in microblogs. Proceedings of the ESWC Workshop, 93-98. Pang, B., & Lee, L. (2008). "Opinion mining and sentiment analysis." Foundations and Trends in Information Retrieval, 2, 1-135. Pavlenko, A. (2008). Emotion and emotion-laden words in the bilingual lexicon. Bilingualism: Language and Cognition, 11(2), 147-164. Pontiki, M., Galanis, D., Pavlopoulos, J., Papageorgiou, H., Androutsopoulos, I., & Manandhar, S. (2014). SemEval-2014 Task 4: Aspect Based Sentiment Analysis. International Workshop on Semantic Evaluation, 27–35, Dublin, Ireland. Association for Computational Linguistics. Smith, A., & Anderson, M. (2018). Social Media Use 2018: Demographics and Statistics. Washington DC: Pew Research Center. Tucker, J. A., Guess, A., Barbera, P., Vaccari, C., Siegel, A., Sanovich, S., Stukal, N., & Nyhan, B. (2018). "Social media, political polarization, and political disinformation: A review of the scientific literature." Political polarization, and political disinformation: a review of the scientific literature (March 19, 2018), Tumasjan A, Sprenger TO, Sandner PG, Welpe IM (2010) Predicting elections with twitter: What 140 characters reveal about political sentiment. Word Journal of The International Linguistic Association, 280, 178–185. Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press. Wiebe, J., Wilson, T., & Cardie, C. (2005). Annotating Expressions of Opinions and Emotions in Language. Language Resources and Evaluation, 39, 165-210.