从自主性到自明性：风景园林专业教育的知识图谱本体构建与探讨

From Autonomy to Self-Evidence: Ontology Construction of a Knowledge Graph for Landscape Architecture Education and Its Implications

关键词:  风景园林  知识图谱  本体构建  自主性  自明性  “景观空间-景观表达-设计思维”

Key words:  landscape architecture  knowledge graph  ontology framework  autonomy  self-evidence  triad of landscape space, landscape representation, and design thinking

摘要:

面对风景园林学科知识体系庞杂化与核心模糊化对专业教育的挑战，引入语义工程方法，旨在构建服务于专业学习的风景园林知识图谱本体。采用“理论-专家”驱动方式，提出了以“景观空间-景观表达-设计思维”为顶层结构的“景观三角形”本体框架，以期实现默会知识显性化、碎片知识结构化，并为该本体框架向学习情境的有效迁移提供支撑。以西安建筑科技大学教学实践为例，探讨了该框架嵌入课程体系、引导设计思维、结构化整合多源知识的具体路径与初步成效，展示了其作为“元语言系统”在整合专业认知、对抗学习碎片化等方面的应用潜力。研究成果不仅是一项技术工具，更是一种学科知识体系在方法论层面的自我反思与重构，为强化风景园林学科的自主性与自明性提供了创新性思路。

Abstract:

In the context of profound transformations in knowledge production and learning paradigms, landscape architecture education faces persistent challenges stemming from a blurred disciplinary core and fragmented knowledge structures. These issues significantly affect students' ability to form a coherent understanding of the field. A nationwide survey conducted in April 2024 by the Landscape Architecture Teaching Guidance Sub-Committee revealed a paradoxical dilemma: teachers and students simultaneously feel overwhelmed by the abundance of knowledge points and yet call for additional content. This contradiction indicates that simple addition or reduction of knowledge cannot resolve the structural problem. The rapid development of artificial intelligence has further intensified difficulties in knowledge screening, integration, and reasoning, making it increasingly urgent to reconstruct a clear system of disciplinary autonomy and self-evidence that can support professional thinking and guide innovation. Responding to this need, the present research introduces semantic engineering methods to construct a knowledge graph ontology tailored to landscape architecture education. The study centers on a fundamental question: whether it is possible to build an inherent semantic framework capable of integrating interdisciplinary knowledge while articulating and transmitting disciplinary consensus. In the era of AI-driven transformations, the autonomy of a discipline lies not in closing boundaries but in its capacity to assimilate external knowledge through a stable internal logic. Likewise, disciplinary self-evidence can be made explicit and debatable through formalized ontological modeling. Based on a theory-expert hybrid approach, the research draws on Charles Sanders Peirce's triadic semiotics - object, sign, interpretant - to construct an innovative top-level ontological structure: "Triad of landscape space, landscape representation, and design thinking". Landscape Space denotes the cognitive object of design, Landscape Representation functions as the symbolic mediator that externalizes spatial cognition and design intentions, and Design Thinking forms the generative process through which meaning, interpretation, and design propositions emerge. Together, these three components constitute a dynamic and recursive mechanism that not only encodes disciplinary knowledge but also simulates design reasoning and enables structured semantic associations. The proposed ontology was validated through a three-semester reconstruction offoundational design courses at Xi'an University of Architecture and Technology, implemented with the first cohort of the "five-to-four-year" curriculum reform in 2024. The framework was systematically integrated into multiple courses as a "meta-language system" capable of linking fragmented content into a coherent learning pathway. It translated traditionally implicit and experience-dependent design processes into nine explicit and operable steps, including cognitive diagnosis, goal definition, project planning, thematic conception, overall and detailed design, implementation, operation and maintenance, and impact evaluation. Teaching practice demonstrated that this structure effectively established a unified semantic framework that enhances disciplinary autonomy across the curriculum, clarified design thinking processes to support disciplinary self-evidence, and helped students build stable cognitive structures by integrating experiential, scientific, and technical knowledge within an open system. The ontology framework also proved transferable beyond formal teaching. Students independently employed it in innovative research projects, such as "Knowledge Graph and AI-Driven Landscape Design Process Innovation: A Case Study of Urban Waterfront Spaces". Through this application, students transitioned from passive recipients of knowledge to active constructors of disciplinary meaning. In AI-assisted design scenarios, the framework served as a semantic constraint that ensured generative tools aligned with professional logic rather than producing visually appealing yet conceptually unsupported results, thereby preserving disciplinary subjectivity in the face of technological change. Overall, this research contributes not only a technical tool but also a methodological reflection and reconstruction of the landscape architecture knowledge system. By addressing challenges of knowledge fragmentation and tacit knowledge transmission, it offers a structured, explicit, and expandable pathway for reinforcing both the autonomy and self-evidence of the discipline. These findings provide valuable insights for strengthening disciplinary identity and advancing landscape architecture education in the digital era.