NOTICE OF THESIS EXAM PATRICIA CYNTHIA CHANDRA

---

NOTICE OF THESIS EXAM

 

PATRICIA CYNTHIA CHANDRA

202200010017

 

 

JULY 15, 2026, 13:00 PM

Building C, 7th fl., Room: C706

 

Adviser           : Dr. Engliana

Examiners     : Dr. Anna Marietta da Silva & Ekarina, Ph.D.

 

Title

 A COMPARATIVE ANALYSIS OF HUMAN AND ARTIFICIAL INTELLIGENCE TRANSLATIONS OF FIGURATIVE LANGUAGES IN ‘THREE-BODY PROBLEM’

ABSTRACT

This study focuses on the relay translation scenario of Liu Cixin’s The Three-Body Problem, where the Indonesian translation of the work was produced via its intermediate English translation. Its core inquiry centers on the effective translation of figurative language. The study adopts a mixed-methods approach, drawing on Halliday’s metafunctional discourse analysis. Using Cochran’s formula, 69 samples were extracted from 83 instances of figurative language. The output of one human translator and five mainstream AI translators were compared after thematic coding and classification of the outputs. The study utilized Nida's CAN standard to assess the retention level of cultural context elements. It discovered that significantly different translation subjects demonstrated varied performances. The human translator predominantly employed Dynamic Equivalence with Cultural Substitution and Paraphrase with a High CAN score of 84% of the time. The human translator expertly adjusted the meaning to maintain coherence and emotion. On the other hand, AI / MT engines relied mainly on literal equivalency, mostly choosing Literal Translation in 78 % of the cases. Although large language models performed well, mainstream neural machine translation systems have often made significant contextual mistakes, as they do not understand figurative spatial, kinematic and cultural reasoning. The study concludes that while AI systems are good at structural language transfer, literary translation still calls for human translators as cultural mediators. The human mediator adapts the aesthetic and socio-cultural value of the source text for the target audience.

Keywords: Human versus Machine Translation, Translation Strategies, Figurative Language, Relay Translation, Nida’s CAN Framework.