“机器翻译与译后编辑”模式下翻译能力的重构与评价——面向翻译专业学位(Master of Translation and Interpreting, MTI)教学的过程–产品双维框架构建

Reconfiguring Translation Competence and Assessment in the Era of Machine Translation and Post-Editing—Construction of a Process-Product Dual Framework for MTI Education

摘要: 随着人工智能技术的飞速发展，机器翻译与译后编辑(machine translation and post-editing, MT + PE)已深刻改变语言服务行业的生产模式。然而，传统翻译教学评价体系长期存在“重结果、轻过程”“重语言、轻技术”的结构性缺陷，难以适应人机协作的新常态。为应对这一挑战，本研究基于功能主义翻译理论、认知心理学及社会技术系统理论(socio-technical systems theory, STS)，采用文献计量与对比分析方法，梳理了近十年译后编辑与翻译能力的研究脉络，并依据ISO 18587:2017《翻译服务——机器翻译输出的译后编辑——要求》(Translation services-Post-editing of machine translation output-Requirements)与欧洲翻译硕士(European Master’s in Translation, EMT)能力框架，系统解析了人工翻译与译后编辑在认知加工层面的差异。研究发现，MT + PE语境下的翻译能力已发生范式转移，从单一的双语转换能力演进为涵盖技术素养、风险评估、决策判断与人机交互效率的综合胜任力。现有以产品为导向的评价模式难以有效反映学习者在技术环境下的过程性能力差异，存在明显的效度局限。为此，本研究构建了一个包含产品与过程双维度的教学评价模型，在保留对译文准确性、规范性评价的同时，引入时间效率、工具交互行为、策略选择模式及元认知反思等过程性指标，并提供了相应的权重分配与评分量表。该框架不仅能够更客观、全面地评估MTI学生在技术驱动环境下的真实专业能力，还可通过评价的反拨作用，引导学生从关注“文本打磨”转向重视“项目交付”与“流程优化”，从而为MTI的教学改革与高层次应用型翻译人才培养提供可行的实施路径。

Abstract: In response to the dominance of Machine Translation plus Post-Editing (MT + PE) as a mainstream production model in the language services industry, this study examines the limitations of traditional product-centered translation competence assessment and constructs an evaluation framework adapted to technology-mediated contexts for Master of Translation and Interpreting (MTI) programs. Grounded in functionalist translation theory and the conceptualization of translation as a socio-technical action, this study analyzes how MT + PE workflows reshape the construct of translation competence. By synthesizing literature on post-editing, the ISO 18587 standard (Translation services—Post-editing of machine translation output—Requirements), and the EMT (European Master’s in Translation) competence framework, the research identifies emerging features of translation competence in technology-driven settings. The study reveals that translation competence in MT + PE contexts exhibits distinct process-oriented and service-oriented characteristics—such as efficiency management and risk control—which traditional product-only assessment fails to capture. Consequently, a dual-dimensional assessment framework is proposed, integrating product indicators (accuracy, consistency, compliance) with process indicators (time efficiency, strategy patterns, tool/QA usage, and decision justification). The proposed process-product dual assessment framework enables a more holistic evaluation of learners’ integrated competencies under human-machine collaboration. It serves to bridge the gap between MTI pedagogy and industry realities, providing actionable pathways for reforming translation competence development and assessment in the era of artificial intelligence.