This paper presents a preliminary exploration of an actor-based model for subject indexing, which considers four types of actors: professional indexers, domain experts, casual indexers, and machine algorithms. The paper describes each of the four actors, enumerating differences in approach, training, methodology, priorities, and tools, as well as similarities and historical collaborations between actors. The paper then explores how the actor-based model for subject indexing might serve as a complement to existing models that focus on processes, methods, disciplinary norms, and cultural biases by defining and exploring the following key properties of an actor-based model for subject indexing: 1) actors are the primary drivers of subject indexing work, 2) observing and understanding many types of actors’ processes in real-life situations is as valuable as prescribing correct methods for professional subject indexing, and 3) multiple and different types of actors can perform subject analysis work and subject representation work on the same information objects, and these hybrid (multi-actor) approaches to subject indexing are explicitly supported. These key properties suggest that an actor-based model for subject indexing might open new research opportunities and encourage new hybrid and collaborative approaches to knowledge organization.