THE ANATOMY OF DIGITAL HATE: A THREE-DIMENSIONAL COMPUTATIONAL ANALYSIS OF ANTI-IMMIGRANT RHETORIC ON THE X PLATFORM

Öz

Social media discourse targeting Syrian refugees in Turkey draws attention as an area where social polarization and identity-based conflicts are reproduced in the digital sphere. The problem addressed in this study is that, while this discourse has been examined in the existing literature using qualitative and limited data, its large-scale and dynamic structure on social media has not been sufficiently mapped. The aim of the research is to empirically reveal the processes of polarization and othering in the digital public sphere by mapping the thematic, emotional, and chronological dimensions of discourse on Syrian refugees using a large dataset covering the period from December 2022 to November 2024 on the X (Twitter) platform. In this regard, topic modeling, machine learning-based hate speech analysis, and multi-label sentiment analysis were applied to Turkish posts collected using keywords such as “migration,” “migrant,” “refugee,” and “Syrian.” The findings reveal two main frameworks shaped around the Syria-migrant axis, which forms the thematic core of the discourse: “victimization” and “threat.” The fact that 99.4% of the analyzed content contains hate speech and that the dominant emotions are anger and fear reveals the toxic nature of discussions in the digital sphere. Additionally, it was found that over time, the discourse shifted from foreign policy and humanitarian aid to internal security and crime themes; while external crises such as earthquakes brought a short-term humanitarian perspective, the discourse quickly returned to a negative framework. The findings were discussed within the framework of Tajfel and Turner's Social Identity Theory, leading to the conclusion that the “us-them” distinction and perceived threats fuel hate speech. In this regard, the study also empirically demonstrates that digital platforms can become echo chambers that exacerbate social polarization.

Anahtar Kelimeler

• Social Media
• Migration
• Text Mining
• Social Identity Theory
• Discourse Analysis

DİJİTAL NEFRETİN ANATOMİSİ: X PLATFORMUNDAKİ GÖÇMEN KARŞITI SÖYLEMİN ÜÇ BOYUTLU HESAPLANMALI ANALİZİ

Öz

Türkiye’de Suriyeli göçmenlere yönelik sosyal medya söylemleri, toplumsal kutuplaşmanın ve kimlik temelli çatışmaların dijital düzlemde yeniden üretildiği bir alan olarak dikkat çekmektedir. Çalışmanın problemi, mevcut literatürde genellikle nitel ve sınırlı veriyle incelenen bu söylemlerin, sosyal medyada geniş ölçekli ve dinamik yapısının yeterince haritalanamamış olmasıdır. Araştırmanın amacı, X (Twitter) platformunda Aralık 2022-Kasım 2024 dönemini kapsayan büyük bir veri seti üzerinde, Suriyeli göçmenlere dair söylemlerin tematik, duygusal ve kronolojik haritasını çıkararak, dijital kamusal alandaki kutuplaşma ve ötekileştirme süreçlerini ampirik olarak ortaya koymaktır. Bu doğrultuda, “göç”, “göçmen”, “sığınmacı”, “Suriyeli” gibi anahtar kelimelerle toplanan Türkçe paylaşımlar üzerinde konu modellemesi, makine öğrenmesi tabanlı nefret söylemi analizi ve çok etiketli duygu analizi uygulanmıştır. Bulgular, söylemin tematik çekirdeğini oluşturan Suriye-göçmen ekseni etrafında şekillenen iki ana çerçeveyi ortaya koymuştur: “Mağduriyet” ve “Tehdit”. Analiz edilen içeriklerin %99,4’ünün nefret söylemi içermesi ve baskın duyguların öfke ile korku olması, dijital alandaki tartışmaların toksik yapısını ortaya koymaktadır. Ayrıca, söylemin zamanla dış politika ve insani yardımdan, iç güvenlik ve suç temalarına evrildiği; deprem gibi dışsal krizlerin ise kısa süreli insani bakış açısı getirse de söylemin hızla olumsuz çerçeveye döndüğü tespit edilmiştir. Bulgular, Tajfel ve Turner’ın Sosyal Kimlik Kuramı çerçevesinde tartışılarak “biz-onlar” ayrımının ve algılanan tehditlerin nefret söylemini beslediği sonucuna ulaşılmıştır. Bu yönüyle çalışma, dijital platformların toplumsal kutuplaşmayı şiddetlendiren yankı odaları haline gelebileceğini de ampirik olarak ortaya koymaktadır.

Anahtar Kelimeler

• Sosyal Medya
• Göç
• Metin Madenciliği
• Sosyal Kimlik Kuramı
• Söylem Analizi

Kaynakça

1. Amaro, A., & Bacao, F. (2024). Topic modeling: A consistent framework for comparative studies. Emerging Science Journal, 8(1), 125–139.
2. Anand, S., Devulapally, N. K., Bhattacharjee, S. D., & Yuan, J. (2023, October). Multi-label emotion analysis in conversation via multimodal knowledge distillation. In Proceedings of the 31st ACM International Conference on Multimedia (pp. 6090–6100). ACM.
3. Arshad, M. U., & Shahzad, W. (2024). Understanding hate speech: The HateInsights dataset and model interpretability. PeerJ Computer Science, 10, e2372.
4. Baden, C., Kligler-Vilenchik, N., & Yarchi, M. (2020). Hybrid content analysis: Toward a strategy for the theory-driven, computer-assisted classification of large text corpora. Communication Methods and Measures, 14(2), 81–101.
5. Bozdag, U. (2024). Framing displaced persons: An analysis of Turkish media’s use of migration metaphors on Twitter. Intersections. East European Journal of Society and Politics, 10(1).
6. Bozdağ, Ç. (2019). Bottom-up nationalism and discrimination on social media: An analysis of the citizenship debate about refugees in Turkey. European Journal of Cultural Studies, 23(5), 712–730.
7. Breuer, J., & Elson, M. (2017). Frustration-aggression theory. In P. Sturmey (Ed.), The Wiley handbook of violence and aggression (pp. 1–12). Wiley Blackwell.
8. Brown, A. (2017). What is hate speech? Part 1: The myth of hate. Law and Philosophy, 36(4), 419–468.

9. Çetiner, S. (2020). Uluslararası toplumun kitlesel göçlere tepkisi ve uluslararası göç yönetimi: Suriye krizi örneği. Mustafa Kemal Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 17(46), 653–672.
10. Ekinci, E., & Omurca, S. I. (2020). NET-LDA: A novel topic modeling method based on semantic document similarity. Turkish Journal of Electrical Engineering and Computer Sciences, 28(4), 2244–2260.
11. Ekman, P. (1992). An argument for basic emotions. Cognition & Emotion, 6(3–4), 169–200.
12. Erdoğan-Ozturk, Y., & Isik-Guler, H. (2020). Discourses of exclusion on Twitter in the Turkish context: #ülkemdesuriyeliistemiyorum (#idontwantsyriansinmycountry). Discourse, Context & Media, 36, 100400.
13. Erdoğan-Öztürk, Y., & Işık-Güler, H. (2021, June 27–July 2). Online hostility at the nexus of migration and gender: Exploring the gendered hate discourse towards Syrian refugee women in Turkey [Conference presentation]. 17th International Pragmatics Conference (IPrA), Winterthur, Switzerland.
14. Erdoğan-Öztürk, Y., & Işık-Güler, H. (2023). The gift of hospitality and the (un)welcoming of Syrian migrants in Turkey. Journal of Immigrant & Refugee Studies, 21(2), 220–234.
15. Farné, A., Ginesta, X., & Paniagua-Rojano, F. J. (2024). Scientific production of open access articles on hate speech: A scoping review. Revista General de Información y Documentación, 34(1), 13–39.
16. Frydenlund, E., Yilmaz Sener, M., Gore, R., Boshuijzen-van Burken, C., Bozdag, E., & De Kock, C. (2023). Out of sight, out of mind: The impact of lockdown measures on sentiment towards refugees. Journal of Information Technology & Politics, 21(6), 1–10.
17. Georgiou, M., & Zaborowski, R. (2017). Media coverage of the “refugee crisis”: A cross-European perspective. Council of Europe.
18. Habermas, J. (1997). Kamusallığın yapısal dönüşümü (T. Bora & M. Sancar, Çev.). İletişim Yayınları
19. Han, Y. (2021). Research on the anti-silence spiral phenomenon in the transmission of internet public opinion—Take the incident of “corporal punishment of a girl with asthma in Guangzhou to cause hematemesis” as an example. Open Journal of Social Sciences, 9(7), 359–366.
20. Hatipoğlu, E., Gökçe, O. Z., Arın, İ., & Saygın, Y. (2019). Automated text analysis and international relations: The introduction and application of a novel technique for Twitter. All Azimuth: A Journal of Foreign Policy and Peace, 8(1), 55–77.
21. Heidenreich, T., Lind, F., & Eberl, J. M. (2020). Computational social science in migration research. Journal of Ethnic and Migration Studies, 46(1), 1–19.
22. Kalogeropoulos, A., Papagiannopoulos, P., Karypidis, P., & Vryzas, N. (2021). A web interface for analyzing hate speech. Future Internet, 13(3), 80.
23. Kostarella, I., Theodosiadou, S., & Touri, M. (2024). Hate speech against journalists: A qualitative study of Greek journalists’ perceptions and experiences. Journalism. Advance online publication.
24. Kurt, G. (2019). Yeni medyada nefret söylemi: YouTube’da Suriyeli mültecilere karşı üretilen nefret söylemi üzerine bir araştırma. The Journal of International Lingual Social and Educational Sciences, 5(1), 1–20.
25. Liu, X., Shi, T., Zhou, G., Liu, M., Yin, Z., Yin, L., & Zheng, W. (2023). Emotion classification for short texts: An improved multi-label method. Humanities and Social Sciences Communications, 10(1), 1–9.
26. Manap, Ç. (2025). Sosyal medyada mülteci temsilinin söylem analizi: Ekşi Sözlük örneği. Pamukkale Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, (67), 83–99.
27. Matthes, J., & Schmuck, D. (2017). The effects of anti-immigrant right-wing populist ads on implicit and explicit attitudes: A moderated mediation model. Communication Research, 44(4), 556–581.
28. Mete, M. (2023). Sosyal kimlik teorisi ve göçmen kimliğinin ötekileştirilmesi süreci: Stereotipleştirmeye yönelik psikopolitik bir analiz. Akademik İncelemeler Dergisi, 18(2), 470–489.
29. Mou, G., & Lee, K. (2021, December). An effective, robust and fairness-aware hate speech detection framework. In 2021 IEEE International Conference on Big Data (Big Data) (pp. 687–697). IEEE.
30. Özdemir, F., & Öner-Özkan, B. (2016). Türkiye’de sosyal medya kullanıcılarının Suriyeli mültecilere ilişkin sosyal temsilleri. Nesne Psikolojisi Dergisi, 4(8), 227–245.
31. Păiș, V. (2024, March). RACAI at ClimateActivism 2024: Improving detection of hate speech by extending LLM predictions with handcrafted features. In Proceedings of the 7th Workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text (CASE 2024) (pp. 67–72). ACL.
32. Parlak, İ., Çakın, Ö., & Kaya, S. (2022). Sosyal medyada sığınmacı algısı: Suriyeli sığınmacıların Türkçe Twitter hesaplarında görünümü. Gümüşhane Üniversitesi İletişim Fakültesi Elektronik Dergisi, 10(2), 948–983.
33. Seate, A. A., & Mastro, D. (2017). Media’s influence on immigration attitudes: An intergroup threat theory approach. Communication Monographs, 84(2), 223–243.
34. Su, Y., Akin, S., & Fuhse, J. A. (2017). A hybrid approach to content analysis: Combining computational and manual methods. Sociological Methodology, 47(1), 1–38.
35. Uludoğan, G., Yüksel, A. E., Tunçer, Ü., Işık, B., Korkmaz, Y., Akar, D., & Özgür, A. (2024, March). Detecting hate speech in Turkish print media: A corpus and a hybrid approach with target-oriented linguistic knowledge. In Proceedings of the 7th Workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text (CASE 2024) (pp. 205–214). ACL.
36. United Nations. (2019). United Nations strategy and plan of action on hate speech. https://rb.gy/3d4r9e United Nations. (2024). Hate speech. https://www.un.org/en/hate-speech
37. Wicaksono, D., Rozaq, H. A. A., & Boz, N. (2025). Emotion recognition for low-resource Turkish: Fine-tuning BERTurk on TREMO and testing on xenophobic political discourse. arXiv. https://arxiv.org/abs/2505.12160
38. Yılmaz, F., Elmas, T., & Eröz, B. (2023). Twitter-based analysis of anti-refugee discourses in Türkiye. Discourse & Communication, 17(3), 298–318.