Abstract: Katharina Reiss’s text typology theory has deeply influenced the translation industry both domestically and internationally since its inception. Its value is not only reflected in providing reliable reference paradigms for the inheritance and innovation of modern and contemporary translation theories but also in its practical application to specialized technical domains. This study bridges the gap between classical linguistic theory and modern technical communication by applying Reiss’s framework to the translation of a complex PhD dissertation on AI-driven football video analysis.
Introduction
Technical translation is frequently misunderstood as a simple process of word-for-word substitution. However, highly specialized scientific texts—such as a dissertation on computer vision—require a sophisticated understanding of "text function." This research addresses two primary challenges: mapping the evolution of Reiss’s 1970s-era theory into the modern era and tackling the inherent rigidities and "hallucinations" of standard machine translation (MT) tools when dealing with informative technical content.
Methodology and Bibliometric Analysis
The researchers employed a dual-method approach to evaluate the current state of the field. First, a bibliometric analysis was conducted using tools like CiteSpace and CNKI to visualize literature from the Web of Science (WOS) and the China National Knowledge Infrastructure. This involved analyzing keyword co-occurrence and clustering to identify research hotspots.
Second, the authors applied Reiss’s functionalist framework—which categorizes texts as informative, expressive, or operative—to a real-world translation project. Since a PhD thesis is primarily informative, the priority was placed on "conceptual content" and "functional equivalence" over literal matching. Finally, a comparative error analysis was performed to contrast raw MT outputs against a human-revised text (RT).
Research Hotspots and Interdisciplinary Trends
A significant finding from the bibliometric data is the emergence of "Machine Translation plus Post-Editing" (MTPE) as a top research hotspot since 2020. This suggests that Reiss’s theory is increasingly serving as a framework for human editors to correct AI-generated translations. Furthermore, the analysis revealed that translation theory is no longer siloed in linguistics; it is now heavily clustered with technical terms like "machine learning," "natural language processing," and "event extraction."
Practical Application: Correcting Technical Logic
In technical translation, the "informative" function is paramount. The study highlights how general-purpose AI often fails at domain-specific nuances. For example, translating the Chinese term for a mathematical "pathological problem" literally as "pathological" (a medical term) instead of the standard engineering term "ill-conditioned."
By prioritizing function over form, translators can ensure technical logic remains intact. This often requires structural changes, such as altering sentence hierarchy or punctuation, to maintain clarity that MT tools like DeepL or Bing often miss.
Results and Domestic Impact
The study identified that while theoretical papers on Reiss are fewer in core Chinese journals, the theory remains a high-utility practical guide, particularly for graduate-level researchers. The authors successfully identified and corrected six categories of errors in MT outputs: improper collocation, missing subjects, logical inconsistencies, mistranslated technical terms, syntax hierarchy errors, and ambiguous pronoun references.
Conclusion
The "Machine Translation plus Post-Editing" (MTPE) model appears to be the inevitable future of the industry. While AI provides a baseline, Reiss’s Text Typology Theory provides the necessary rules for human editors to refine these outputs into professional-grade technical documents. Future work should focus on standardizing these post-editing rules to further integrate functionalist theory with automated workflows.
Summary updated