Song Translation and ChatGPT

Alaz Pesen

Makale detayları

Song Translation and ChatGPT

Yazarlar:

Alaz Pesen

Yayıncı:

Bolu Abant Izzet Baysal University

Ana Başlıklar:

Translation and Interpretation Studies

Dergi:

Çeviribilim Dergisi Vol.2 Issue 1, pp.1-16

Belge Tipi:

Makale

Yayın Tarihi:

2/2024

Belge Dili:

Türkçe

Document ID (Org No.):

2001312598

pISSN:

3023-4344

Anahtar kelimeler:

Özet:

 Since the artificial intelligence chatbot ChatGPT, written in the Python programming language by the OpenAI company, entered our lives on November 30, 2022, it has been used in almost every field. This study is written to understand how ChatGPT can converge with the field of song translation, a very interesting and widely practiced form of translation that attracted the attention of translation studies only for the past quarter century. The main question that the study seeks to answer is whether ChatGPT can translate songs. Song translation is a translation genre with its own subfields. Translating a song sung or staged in a series into the target language as subtitles can be defined as song translation. Similarly, translating a song so that it can be sung in the target language is another subfield of song translation. In this subfield, there are also sub-areas such as meaning-focused song translation, sound-focused phonemic song translation, as well as writing entirely new lyrics in the target language. The theoretical framework of the study is based on the tetrathlon approach based on Peter Lows pentathlon approach: ChatGPT 3.5 has been asked to perform translations focused on meaning, syllable count, rhyme, and localization in sequence. The song Je Veux has been selected since it is replete with cultural-specific elements, a real challenge for all song translators. Turkish translations made by ChatGPT at each step are examined, and its final translation is compared with the translation made by a real person in the last step. The innovative tetrathlon approach aims to contribute to the intersection of descriptive song translation studies and machine translation.

 Since the artificial intelligence chatbot ChatGPT, written in the Python programming language by the OpenAI company, entered our lives on November 30, 2022, it has been used in almost every field. This study is written to understand how ChatGPT can converge with the field of song translation, a very interesting and widely practiced form of translation that attracted the attention of translation studies only for the past quarter century. The main question that the study seeks to answer is whether ChatGPT can translate songs. Song translation is a translation genre with its own subfields. Translating a song sung or staged in a series into the target language as subtitles can be defined as song translation. Similarly, translating a song so that it can be sung in the target language is another subfield of song translation. In this subfield, there are also sub-areas such as meaning-focused song translation, sound-focused phonemic song translation, as well as writing entirely new lyrics in the target language. The theoretical framework of the study is based on the tetrathlon approach based on Peter Lows pentathlon approach: ChatGPT 3.5 has been asked to perform translations focused on meaning, syllable count, rhyme, and localization in sequence. The song Je Veux has been selected since it is replete with cultural-specific elements, a real challenge for all song translators. Turkish translations made by ChatGPT at each step are examined, and its final translation is compared with the translation made by a real person in the last step. The innovative tetrathlon approach aims to contribute to the intersection of descriptive song translation studies and machine translation.

 Since the artificial intelligence chatbot ChatGPT, written in the Python programming language by the OpenAI company, entered our lives on November 30, 2022, it has been used in almost every field. This study is written to understand how ChatGPT can converge with the field of song translation, a very interesting and widely practiced form of translation that attracted the attention of translation studies only for the past quarter century. The main question that the study seeks to answer is whether ChatGPT can translate songs. Song translation is a translation genre with its own subfields. Translating a song sung or staged in a series into the target language as subtitles can be defined as song translation. Similarly, translating a song so that it can be sung in the target language is another subfield of song translation. In this subfield, there are also sub-areas such as meaning-focused song translation, sound-focused phonemic song translation, as well as writing entirely new lyrics in the target language. The theoretical framework of the study is based on the tetrathlon approach based on Peter Lows pentathlon approach: ChatGPT 3.5 has been asked to perform translations focused on meaning, syllable count, rhyme, and localization in sequence. The song Je Veux has been selected since it is replete with cultural-specific elements, a real challenge for all song translators. Turkish translations made by ChatGPT at each step are examined, and its final translation is compared with the translation made by a real person in the last step. The innovative tetrathlon approach aims to contribute to the intersection of descriptive song translation studies and machine translation.

 OpenAI şirketinin Python programlama diliyle yazdığı yapay zekâ sohbet robotu ChatGPT 30 Kasım 2022de hayatlarımıza girdiğinden beri, hemen her alanda kullanılmaya başlanmıştır. Farklı dil ve kültürlerle yalnızca bir mesaj ya da e-posta aracılığıyla iletişime geçebildiğimiz şu günlerde, ChatGPTnin en yaygın kullanıldığı alanların başında çevirinin geldiğini söylemek de herhalde şaşırtıcı olmayacaktır. Elinizdeki çalışma çevirinin çok ilgi çekici, son derece yaygın, ancak çeviribilimin ilgisini yalnızca çeyrek yüzyıldır çekmekte olan bir alanıyla, yani şarkı çevirisiyle, ChatGPTnin nasıl bir ortak paydada buluşabileceğini anlamak amacıyla kaleme alınmıştır. Çalışmanın cevabını aradığı başlıca soru şudur: ChatGPT şarkı çevirebilir mi? Şarkı çevirisi, bilindiği gibi, kendi alt alanlarına sahip bir çeviri türüdür. Bir dizide söylenen şarkının ya da sahnelenen operanın erek dilde alt ya da üst yazı olarak çevrilmesi şarkı çevirisi olarak tanımlanabileceği gibi, A dilinde söylenen bir şarkının B diline de söylenebilir şekilde çevrilmesi yine şarkı çevirisinin bir alt alanıdır. Bu alt alanda da anlam odaklı, ses odaklı phonemic ve kaynak şarkının melodisi üzerine erek dilde yepyeni sözler yazma gibi alt alanlar bulunmaktadır. Bu son üç alt alana sırasıyla I Will Survive-Bambaşka Biri, Ola Se Thimizioun-Olmasa Mektubun, Cest Ecrit Dans Le Ciel-Bak Bir Varmış Bir Yokmuş kaynak-erek şarkı çiftleri örnek gösterilebilir. Çalışmanın kuramsal çerçevesini Peter Lowun pentatlon yaklaşımından yola çıkarak geliştirdiğim tetratlon yaklaşımı oluşturmaktadır: ChatGPT-3.5tan sırasıyla anlam, hece sayısı, kafiye, ve yerelleştirme odaklı çeviriler yapması istenmiştir. Çalışmada kültüre özgü ögelerin sıklıkla gözlendiği bir şarkı olan Je Veux tercih edilmiş, ChatGPTnin her adımda yaptığı Türkçe çeviriler incelenmiş, en son adımdaki çevirisi ise şarkının insan tarafından yapılmış çevirisiyle karşılaştırılmıştır. Çalışmada yöntem olarak benimsenen ve ilk kez uygulanan tetratlon yaklaşımı makine çevirisi alanında yapılacak betimleyici şarkı çevirisi incelemelerine de katkı sağlamaktadır.

Makale detayları

Song Translation and ChatGPT

Yazarlar:

Alaz Pesen

Yayıncı:

Bolu Abant Izzet Baysal University

Ana Başlıklar:

Translation and Interpretation Studies

Dergi:

Çeviribilim Dergisi Vol.2 Issue 1, pp.1-16

Belge Tipi:

Makale

Yayın Tarihi:

2/2024

Belge Dili:

Türkçe

Document ID (Org No.):

2001312598

pISSN:

3023-4344

Song translation

ChatGPT

OpenAI

Python

Translation Technologies

Translation Studies

AIŞarkı çevirisi

ChatGPT

OpenAI

Python

Çeviri Teknolojileri

Çeviribilim

Yapay Zekâ

Sayı

Makale detayları

Yazarlar: Alaz Pesen

Yayıncı: Bolu Abant Izzet Baysal University

Ana Başlıklar: Translation and Interpretation Studies

Dergi: Çeviribilim Dergisi Vol.2 Issue 1, pp.1-16

Belge Tipi: Makale

Yayın Tarihi: 2/2024

Belge Dili: Türkçe

Document ID (Org No.): 2001312598

pISSN: 3023-4344

Sayı

Yazarlar:

Alaz Pesen

Yayıncı:

Bolu Abant Izzet Baysal University

Ana Başlıklar:

Translation and Interpretation Studies

Dergi:

Çeviribilim Dergisi Vol.2 Issue 1, pp.1-16

Belge Tipi:

Makale

Yayın Tarihi:

2/2024

Belge Dili:

Türkçe

Document ID (Org No.):

2001312598

pISSN:

3023-4344