R0021/2026-03-25/Q006/SRC02/E01¶
RFC 2119 original scope and usage constraints
URL: https://www.rfc-editor.org/rfc/rfc2119.html
Extract¶
RFC 2119 (March 1997, S. Bradner) states these keywords "MUST only be used where it is actually required for interoperation or to limit behavior which has potential for causing harm."
The standard explicitly constrains its scope: keywords "must not be used to try to impose a particular method on implementors where the method is not required for interoperability."
Relevance to Hypotheses¶
| Hypothesis | Relationship | Strength |
|---|---|---|
| H1 | N/A | Provides context but does not confirm or deny AI application |
| H2 | N/A | The original scope does not prohibit AI application |
| H3 | Supports | RFC 2119's original scope (interoperability, harm prevention) aligns with AI behavioral constraints more than with prompt design |
Context¶
RFC 2119 was designed for technical interoperability specifications. Its application to AI behavioral constraints is an extension of the original intent, though the "limit behavior which has potential for causing harm" language is relevant to AI safety specifications.