Yapay Zekâ Tabanli Dil Modellerinde Anlamlandirma Sorunu: Geleneksel Sözlük Tanimlarinin Yeniden Değerlendirilmesi

Öz

Bu çalışmanın amacı, yapay zekâ tabanlı büyük dil modelleriyle geleneksel sözlük tanımları arasındaki farklılıkları ortaya koymak ve bu farklılıkların Türkçe öğretimine yansımalarını değerlendirmektir. Araştırma, tanımların karşılaştırılmasıyla sınırlı kalmamakta; ortaya çıkan bulguların Türkçe öğretimiyle ilişkilendirilmesi ve bu bağlamda geliştirilecek yöntem ve materyallere ışık tutması açısından da önem taşımaktadır. Araştırmada, nitel araştırma yöntemlerinden doküman analizi benimsenmiştir. Güncel Türkçede sıkça karşılaşılan on somut ve on soyut olmak üzere toplam yirmi kelime, amaçlı örnekleme yöntemiyle belirlenmiştir. Kelimelerin tanımı, Türk Dil Kurumu (TDK) çevrim içi Güncel Sözlükten ve ayrıntılı yönergelerle yönlendirilen GPT-4 büyük dil modelinden elde edilmiştir. Tanımlar, etimolojik köken, anlam çeşitliliği, birleşik yapılar, deyimler, örnek cümleler ile teknik ve kültürel kullanımlar bakımından tablolaştırılmış; elde edilen veriler içerik analiziyle kodlanmış, ikinci bir uzman tarafından güvenilirlik açısından değerlendirilmiştir. Araştırma bulgularına göre; somut kavramlarda her iki kaynağın benzer tanımlar sunduğu, soyut ve çok karmaşık kavramlarda ise GPT-4 modelinin bağlamsal derinlik ve örnek çeşitliliği açısından öne çıktığı belirlenmiştir.

Anahtar Kelimeler

• Sözlükbilim
• Yapay Zekâ
• Türkçe Öğretimi
• Somut ve Soyut Kavramlar
• Karşılaştırmalı Analiz

The Problem of Understanding in Artificial Intelligence-Based Language Models: Re-evaluation of Traditional Dictionary Definitions

Abstract

The aim of this study is to reveal the differences between artificial intelligence-based large language models and traditional dictionary definitions, and to evaluate how these differences reflect on Turkish language teaching. The research is not limited to the comparison of definitions; it also emphasizes the importance of associating the findings with Turkish language instruction and shedding light on the development of methods and materials in this context. In the study, document analysis—one of the qualitative research methods—was adopted. Twenty words, ten concrete and ten abstract, that are frequently encountered in contemporary Turkish were selected using purposive samp-ling. The definitions of these words were obtained from the Turkish Language Association’s (TDK) online contemporary dictionary and from the GPT-4 large language model, guided by detailed instructions. The definitions were tabulated in terms of etymological origin, semantic variety, compound structures, idioms, example sentences, and technical or cultural usage; the collected data were coded through content analysis and evaluated for reliability by a second expert. According to the research findings, while both sources provided similar definitions for concrete concepts, the GPT-4 model stood out in terms of contextual depth and diversity of examples, especially for abstract and multi-layered concepts.

Keywords

• Lexicology, Artificial Intelligence, Turkish Language Teaching, Concrete and Abstract Concepts, Comparative Analysis

Kaynakça

1. Abel, A., & Meyer, C. M. (2013). The dynamics outside the paper: User contributions to online dictionaries. Lexikos, 23(1), 1–22. https://doi.org/10.5788/23-1-1235
2. Almeman, F. Y. (2024). Dictionary examples in the era of large language models. In Proceedings of the LREC-COLING 2024 Conference (pp. 13664–13673). European Language Resources Associa-tion. https://aclanthology.org/2024.lrec-main.1366/
3. Almeman, F., Schockaert, S., & Espinosa-Anke, L. (2024). WordNet un-der scrutiny: Dictionary examples in the era of large language models. In Proceedings of LREC-COLING 2024 (pp. 17683–17695). European Language Resources Association. https://aclanthology.org/2024.lrec-main.1538.pdf
4. Atkins, B. T. S., & Rundell, M. (2008). The Oxford guide to practical lexicography. Oxford University Press.
5. Bański, P., & Przepiórkowski, A. (2009). Stand-off annotation: The case for flexible and extensible annotation. In Proceedings of the Linguistic Annotation Workshop (pp. 21–28). Association for Computational Linguistics.
6. Bats, T., Kallas, J., & Kilgarriff, A. (2022). Multilingual examples in mo-dern e-lexicography. In Proceedings of the XVIII EURALEX International Cong-ress (pp. 280–284). EURALEX.
7. Béjoint, H. (2010). The lexicography of English: From origins to present. Oxford University Press.
8. de Schryver, G. M. (2023). Generative AI and lexicography: The current state of the art. Lexikos, 33, 1–44. https://doi.org/10.5788/33-1-1780

9. Dziemianko, A., & Lew, R. (2016). Electronic dictionary entries with ani-mated pictures: More than just fun? Lexicographica, 32(1), 245–267. https://doi.org/10.1515/lexi-2016-0011
10. Fontenelle, T. (2013). Practical lexicography: A reader. Oxford Univer-sity Press.
11. Fuertes-Olivera, P. A. (2024). ChatGPT, prompting and some lexicograp-hical tasks. In Current perspectives on lexicographical practices in the era of AI and ChatGPT (pp. 13–29). Peter Lang.
12. Landau, S. I. (2001). Dictionaries: The art and craft of lexicography (2nd ed.). Cambridge University Press.
13. Lew, R. (2023). ChatGPT as a COBUILD lexicographer. In Proceedings of the XX EURALEX International Congress (pp. 539–550). EURA-LEX. https://euralex.org/publications/chatgpt-as-a-cobuild-lexicographer/
14. Lew, R. (2024a). Dictionaries and lexicography in the AI era. International Journal of Lexicography. Advance online publica-tion. https://doi.org/10.1093/ijl/ecae001
15. Lew, R. (2024b). The effectiveness of ChatGPT as a lexical tool for Eng-lish: Insights from mono- and bilingual reference sources. In Proceedings of Eu-ralex 2024. (Basımda).
16. McCarthy, D. (2023). AI and the future of dictionary examples. In eLex 2023: Proceedings of the 7th eLex Conference (pp. 145–157). Lexical Computing.
17. OpenAI. (2024). ChatGPT (GPT-4o) [Büyük dil modeli]. https://chat.openai.com/ (Erişim tarihi: 02 Mayıs 2025).
18. Phoodai, C., Rikk, R., Medveď, M., Měchura, M., Kosem, I., Kallas, J., Tiberius, C., & Jakubíček, M. (2023). Exploring the capabilities of ChatGPT for lexicographical purposes: A comparison with the Oxford Advanced Learner’s Dictionary within the microstructural framework. In Proceedings of eLex 2023 (pp. 335–365). https://elex.link/elex2023/wp-content/uploads/77.pdf
19. Rees, G. P., & Lew, R. (2023). The effectiveness of OpenAI GPT-generated definitions versus definitions from an English learners’ dictionary in a lexically orientated reading task. International Journal of Lexicography. Advance online publication. https://doi.org/10.1093/ijl/ecad030Reviriego,
20. P., Martínez, C., Pérez, J. A., Miranda, G., & Ronzano, F. (2024). Comparing the vocabulary and lexical diversity of ChatGPT and humans. In Proceedings of the 17th International Conference on Language Resources and Evaluation (LREC-COLING 2024) (pp. 17610–17617). European Language Re-sources Association. https://aclanthology.org/2024.lrec-main.1529.pdf
21. Rundell, M., & Kilgarriff, A. (2011). Automating the creation of dictiona-ries: Where will it all end? In S. Granger & M. Paquot (Eds.), Electronic lexicog-raphy (pp. 257–281). Oxford University Press.
22. Sinclair, J. (1991). Corpus, concordance, collocation. Oxford University Press.
23. Svensén, B. (2009). A handbook of lexicography: The theory and practice of dictionary-making. Cambridge University Press.
24. The Asian Association for Lexicography. (2023). Proceedings of the 17th ASIALEX International Conference: Lexicography in Asia – Challenges and Pros-pects in the Digital Era (July 21–23, 2023). Yonsei University, Seoul, Ko-rea. http://asialex2023.kr/proceedings.html
25. Tummers, J. (2022). Lexicography in a digital age. International Journal of Lexicography, 35(4), 472–490. https://doi.org/10.1093/ijl/ecab035
26. Türk Dil Kurumu. (t.y.). Güncel Türkçe Söz-lük. https://sozluk.gov.tr/ (Erişim tarihi: 02 Mayıs 2025).
27. Wiegand, H. E. (1998). Wörterbuchforschung: Untersuchungen zur Wör-terbuchbenutzung, zur Theorie, Geschichte, Kritik und Automatisierung der Lexikographie. Max Niemeyer Verlag.
28. Zgusta, L. (1971). Manual of lexicography. Mouton.