Peirce's triadic sign — object, representamen, interpretant — explains Digital Humanities' AI test: meaning lives in the interpretant, with the analyst, not the model.
#DigitalHumanities #AI #LLM #ComputationalHumanities
The most instructive intersection in current Digital Humanities is Peircean semiotics — because it names precisely what generative AI places at risk. Computational modelling is best understood as a triadic sign process: object, representamen, interpretant. Meaning does not inhere in the model; it is produced in the interpretant, the act of interpretation performed by a reader, not an instrument.
This becomes concrete when interpreted at the level of the knowledge graph. RDF encodes a dyadic relation; it cannot natively express the irreducible third term of a sign. The interpretant must therefore be modelled explicitly — via provenance annotations, citation binding, or interpretive metadata — or it is silently lost. Provenance-aware retrieval and evidence-grounded agentic scholarship (SPIRE, Chronos) are, under this reading, attempts to keep the interpretant legible within the computational pipeline.
The humanities are where AI confronts its most rigorous test: not whether models can perform, but whether interpretation survives them.