The paper introduces the framework of update semantics, which redefines the meaning of a sentence as an operation on information states rather than a specification of truth conditions. An update system is defined by a set of relevant information states, Σ, and a function [ ] that assigns to each sentence φ an operation [φ] on Σ. The result of updating state σ with sentence φ is denoted as σ[φ]. The paper emphasizes that update semantics does not always add the informational content of φ to the existing information state σ, and discusses constraints that must be met for an update system to be additive. These constraints include the existence of a minimal state, the definition of σ[φ] for every σ, and idempotence, where σ[φ] should be defined only if σ already contains the information φ provides. The paper also explores the concept of acceptance, where φ is accepted in σ if σ[φ] = σ, indicating that φ's information is already subsumed by σ.The paper introduces the framework of update semantics, which redefines the meaning of a sentence as an operation on information states rather than a specification of truth conditions. An update system is defined by a set of relevant information states, Σ, and a function [ ] that assigns to each sentence φ an operation [φ] on Σ. The result of updating state σ with sentence φ is denoted as σ[φ]. The paper emphasizes that update semantics does not always add the informational content of φ to the existing information state σ, and discusses constraints that must be met for an update system to be additive. These constraints include the existence of a minimal state, the definition of σ[φ] for every σ, and idempotence, where σ[φ] should be defined only if σ already contains the information φ provides. The paper also explores the concept of acceptance, where φ is accepted in σ if σ[φ] = σ, indicating that φ's information is already subsumed by σ.