Despite unprecedented growth in biodiversity data, a persistent gap remains between what is known and what is acted upon. Existing frameworks such as the FAIR and CLEAR Principles have improved data accessibility and interpretability but do not provide the components required to translate knowledge into context-sensitive action. We argue that closing this knowledge-action gap requires a shift toward statement-centred and action-oriented knowledge infrastructures. We identify a fundamental distinction between actionability as the structural capacity of a representation to support operations and applicability as the epistemic validity of using that knowledge in a specific context. Building on the Semantic Units Framework, we introduce Action Units as structured extensions of plan specifications that make applicability conditions and contextual grounding explicit as first-class typed components. Three types are distinguished, epistemic, transformational, and intervention action units, corresponding to three operation classes that define a minimal operational architecture for actionable knowledge. Action units can also be granularly composed across operation classes, reflecting the cross-class character of real-world knowledge-driven processes. Conditional action units, operationalized as executable IF-THEN structures, enable knowledge graphs to function as graph-native decision-support systems, constituting a transition toward post-FAIR knowledge infrastructures. Applied to biodiversity science, the framework reinterprets documented intervention and epistemic failures as consequences of incomplete action unit structures and constructs worked examples across all three operation classes. We propose the TripleA Principle: Actionability, Applicability, and Auditability, as a guiding framework for next-generation knowledge infrastructure design extending the FAIR and CLEAR Principles.

翻译：尽管生物多样性数据呈空前增长态势，但已知内容与实际行动之间始终存在显著鸿沟。现有框架如FAIR原则与CLEAR原则虽提升了数据的可访问性与可解释性，却未提供将知识转化为情境敏感行动所需的组件。我们认为，弥合这一知识-行动鸿沟需要转向以陈述为中心、行动导向的知识基础设施。我们识别出"可操作"作为表征支持运算的结构能力与"可应用"作为特定情境中使用该知识的认知有效性之间的根本区别。基于语义单元框架，我们引入行动单元作为计划规范的结构化扩展，将适用性条件与情境依据明确视为头等类型化组件。行动单元分为三种类型：认知型、转化型与干预型行动单元，对应定义可操作知识最小运算架构的三类运算范式。行动单元还可跨运算类别进行粒度组合，反映真实世界知识驱动过程的跨类特性。条件型行动单元通过可执行的IF-THEN结构实现，使知识图谱能够充当图原生决策支持系统，构成面向后FAIR知识基础设施的转型。在生物多样性科学应用中，该框架将已记录的干预失败与认知失败重新诠释为行动单元结构不完整的后果，并构建覆盖全部三类运算范式的示范案例。我们提出三重A原则：可操作性、可应用性与可审计性，作为扩展FAIR与CLEAR原则的下一代知识基础设施设计指导框架。