How did Rudolf Carnap later revise his early theory of reducibility?

Rudolf Carnap’s early theory of reducibility, as presented in his 1928 work The Logical Structure of the World, was quite narrow. It asserted that all terms describing empirical facts could be fully defined by, and thus fully translated into, terms referring exclusively to immediate experience.1

Carnap later revised this into a more liberal version of empiricism, which he detailed in his essay "Testability and Meaning" (1936–37). He moved away from the requirement of full definability, arguing instead that scientific terms could be only partly defined through:

• Reduction sentences: Logically refined versions of operational definitions.1
• Observation sentences: Statements whose truth can be checked via direct observation.1

In this revised framework, Carnap emphasized that empirical tests typically do not provide absolute proof or disproof but rather offer varying degrees of confirmation for a statement. He continued to liberalize these concepts in later writings to better account for the complex relationship between scientific theories and evidence.1