This paper introduces WORDNET-AFFECT, a linguistic resource for lexical representation of affective knowledge. It was developed from WORDNET by selecting and labeling synsets that represent affective concepts. The resource aims to provide a hierarchy of "affective domain labels" independent from the existing domain hierarchy in WORDNET DOMAINS. The development process involved two stages: identifying a core set of affective synsets and extending this core using WORDNET relations. The core was initially manually created in a resource named AFFECT, which contained 1,903 terms related to mental states. The affective information from AFFECT was projected onto WORDNET synsets, and additional manual labeling was performed to assign affective labels (a-labels) to the synsets. The final version of WORDNET-AFFECT includes 2,874 synsets and 4,787 words. The resource is useful in various applications, including affective reasoners, information and tutoring tools, affective text sensing systems, and computational humor. Future work focuses on extending the number of affective synsets, integrating common sense knowledge, and enhancing the organization of a-labels for better communication and persuasion.This paper introduces WORDNET-AFFECT, a linguistic resource for lexical representation of affective knowledge. It was developed from WORDNET by selecting and labeling synsets that represent affective concepts. The resource aims to provide a hierarchy of "affective domain labels" independent from the existing domain hierarchy in WORDNET DOMAINS. The development process involved two stages: identifying a core set of affective synsets and extending this core using WORDNET relations. The core was initially manually created in a resource named AFFECT, which contained 1,903 terms related to mental states. The affective information from AFFECT was projected onto WORDNET synsets, and additional manual labeling was performed to assign affective labels (a-labels) to the synsets. The final version of WORDNET-AFFECT includes 2,874 synsets and 4,787 words. The resource is useful in various applications, including affective reasoners, information and tutoring tools, affective text sensing systems, and computational humor. Future work focuses on extending the number of affective synsets, integrating common sense knowledge, and enhancing the organization of a-labels for better communication and persuasion.