2004Unpublished venueRequires access

Parser Combinators, (Simply) Indexed Grammars, Natural Language Parsing

Jan van Eijck

Open publisher page 0 citations

Abstract

Parser combinators [4, 6] are higher order functions that transform parsers into parsers. Parsing with context free grammars can be handled by defining combinators for the key operations in context free grammar rules: recognizing epsilon, recognizing a terminal, choice between rewrite rules for a given nonterminal, and sequencing under a nonterminal. After explaining and illustrating this in some detail, we demonstrate how adding two extra combinators, push and pop, allows us to define stack parser combinators that can be used for the analysis of movement (or extraction) in natural language. Push and pop increase the expressive power of parser combinators to beyond CF. The corresponding grammar formalism turns out to be a slight variation on that of indexed grammars. Studying indexed grammars from the perspective of deductive parsing, we propose Earley style parsing algorithms for them, and for the variation on them (“simply indexed grammars”) that corresponds to the parser combinators with push and pop.

About this research paper

What this paper is about

Parser combinators [4, 6] are higher order functions that transform parsers into parsers. Parsing with context free grammars can be handled by defining combinators for the key operations in context free grammar rules: recognizing epsilon, recognizing a terminal, choice between rewrite rules for a given nonterminal, and sequencing under a nonterminal. After explaining and illustrating this in some detail, we demonstrate how adding two extra combinators, push and pop, allows us to define stack parser combinators that can be used for the analysis of movement (or extraction) in natural language. Push and pop increase the expressive power of parser combinators to beyond CF. The corresponding grammar formalism turns out to be a slight variation on that of indexed grammars. Studying indexed grammars from the perspective of deductive parsing, we propose Earley style parsing algorithms for them, and for the variation on them (“simply indexed grammars”) that corresponds to the parser combinators with push and pop.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Parser combinators [4, 6] are higher order functions that transform parsers into parsers. Parsing with context free grammars can be handled by defining combinators for the key operations in context free grammar rules: recognizing epsilon, recognizing a terminal, choice between rewrite rules for a given nonterminal, and sequencing under a nonterminal. After explaining and illustrating this in some detail, we demonstrate how adding two extra combinators, push and pop, allows us to define stack parser combinators that can be used for the analysis of movement (or extraction) in natural language. Push and pop increase the expressive power of parser combinators to beyond CF. The corresponding grammar formalism turns out to be a slight variation on that of indexed grammars. Studying indexed grammars from the perspective of deductive parsing, we propose Earley style parsing algorithms for them, and for the variation on them (“simply indexed grammars”) that corresponds to the parser combinators with push and pop.

Key concepts: Parser combinator, Parsing expression grammar, Combinatory logic, Computer science, Programming language, L-attributed grammar, Context-free grammar, Top-down parsing

Related papers

Back to paper searchBrowse research topicsOriginal source
Parser Combinators, (Simply) Indexed Grammars, Natural Language Parsing — Research Paper | ScholarLens