Patent Translation Quality Assessment: A Weighted Multidimensional Framework Adjusted to Scope Preservation and Claim-Description Asymmetry with a Risk Index Indicator System

<title>Abstract</title> <p> The currently popular translation quality assessment (TQA) models prove inadequate for the unique challenges of patent texts. Inappropriate handling in translation can lead to unintended narrowing of claim scope, terminological inconsistency, or added-matter risks, thereby compromising the legal and commercial value of the patent. This paper proposes a weighted multidimensional framework for patent translation quality assessment, drawing on legal translation theory, CNIPA claim interpretation principles, and Multidimensional Quality Metrics (MQM). The framework evaluates translations across seven weighted dimensions, with particular emphasis on claim scope preservation and claim-description differentiation. Through a detailed parallel textual analysis of a high-stakes English-Chinese patent featuring the coined terms “Clutch” and “Boom,” the study demonstrates how the proposed rubric can systematically identify risks related to scope preservation and the maintenance of appropriate abstraction levels between claims and description. The paper also introduces a “Risk Index” system to provide practitioners with a practical overview of potential vulnerabilities. Findings indicate that effective patent translation requires not only linguistic accuracy but a sophisticated understanding of legal function and the distinct rhetorical purposes of different patent sections. The framework offers practical value for translators, patent attorneys, and machine translation post-editing workflows, while laying the groundwork for larger-scale corpus-based research. <bold>MSC Classification</bold> : 91F20, 68T50 <bold>JEL Classification</bold> : K11, O34, O31 </p>