Can Pikachu evolve in Turkish?
STRIDON: Journal of Studies in Translation and Interpreting, cilt.6, sa.1, ss.27-53, 2026 (Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 6 Sayı: 1
- Basım Tarihi: 2026
- Doi Numarası: 10.4312/stridon.6.1.27-53
- Dergi Adı: STRIDON: Journal of Studies in Translation and Interpreting
- Derginin Tarandığı İndeksler: Scopus, Directory of Open Access Journals, European Reference Index for the Humanities and Social Science (ERIH PLUS), MLA International Bibliography
- Sayfa Sayıları: ss.27-53
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- İstanbul Üniversitesi Adresli: Evet
Özet
This study replicates and extends the study of Moreno García and Mangiron (2024) and investigates the efficacy of GPT-5 in transcreating Pokémon names into Turkish within the Prompt-based Conversational Machine Translation (PCMT) framework. By applying a novel typology across five case studies, the research evaluates the model’s performance in creativity, semantic integrity, cultural resonance and phonological plausibility. The findings reveal that while GPT-5 demonstrates high semantic agility and morphological inventiveness, it exhibits vulnerabilities regarding Turkish morphophonological constraints and sociolinguistic safety. These limitations necessitate a mandatory layer of “constrained transcreation”, refuting the notion of fully autonomous AI generation. Consequently, the study suggests that the integration of LLMs does not signal the obsolescence of the human translator, but rather the evolution of their role from a linguistic mediator to a strategic “prompt architect” and a culturally and sociolinguistically sensitive assessor of the AI output. Ultimately, this research validates a symbiotic workflow where computational power is tempered by the indispensable cultural and structural oversight of a human co-agent.