Environment-Sensitive Lexicographic Disambiguation for Contextual Parsing

Abstract

This paper presents a deterministic algorithm for resolving ambiguity in parse trees using a global mutable context. The proposed method applies a tournament-style selection process to competing derivations at each non-terminal, systematically discarding alternatives whose non-terminal subtrees are not selected by the contextual decision mechanism. Unlike approaches that rely on post-processing, the algorithm maintains semantic state throughout incremental Abstract Syntax Tree (AST) building, allowing earlier decisions to influence the resolution of future ambiguities. This context-aware strategy enables consistent and procedural disambiguation after parsing. It features syntactic disambiguation based on a document environment instead of relying on ad-hoc rules, thus able to model complex relationships between previous constructs and the different derivations for a same non-terminal.