Langbda

Langbda is an incremental parser that models how a listener receives a linear token stream and builds structured syntactic interpretations.

Current focus is a controlled English grammar with movement-aware parsing behavior inspired by Minimalist Grammar style dependencies.

What It Can Do Now

• Parse core clause types in a curated grammar:
  • declaratives,
  • PP-attachment ambiguity,
  • do-support declaratives,
  • yes-no questions,
  • wh-object questions.
• Keep ambiguity where intended:
  • the child ate an apple in the room. yields exactly 2 parses.
• Track movement-chain lifecycle events in derivation metadata.
• Render parse trees as PNG and export derivation traces as JSON.
• Use memoized search to avoid infinite recursion on cyclic functional rules.

Quick Start

Build:

cargo build

Run the core fixture batch and export both PNG and JSON:

cargo run -- --batch core --format both --output-dir assets/examples

Parse one sentence:

cargo run -- --sentence "did the child eat an apple?" --target Sentence --format both --output-dir assets/examples

Show CLI help:

cargo run -- --help

CLI Options

langbda [--sentence "..."] [--target Sentence] [--batch core] [--format png|json|both] [--output-dir DIR] [--no-movement-arrows]

• --sentence: parse a single input sentence.
• --target: parse target category (default Sentence).
• --batch core: run built-in core fixtures.
• --format: artifact format (png, json, or both).
• --output-dir: output directory for artifacts.
• --no-movement-arrows: disable movement arrows in rendered PNG trees.

Output Artifacts

For each parse, the CLI writes deterministic files:

• {sentence_slug}__{target_slug}__parse-XX.png
• {sentence_slug}__{target_slug}__parse-XX.json

JSON includes:

• sentence and target,
• token stream,
• ordered derivation steps (merge/move/check),
• movement chain lifecycle events,
• final well-formedness and unresolved chain diagnostics.

Example Ambiguity

The model captures two interpretations of:

The child ate an apple in the room.

Grammar and Resources

• Example lexicon: assets/lexicons/en.lexicon
• Capability contract: docs/capability_matrix.md
• Quality notes: docs/quality_report.md

Limits and Scope

• The English lexicon is intentionally small and diagnostic-oriented.
• This is not an open-domain English parser.
• Some question analyses still rely on a constrained normalization bridge in the interpreter, so strict surface-faithful derivation for all constructions is still in progress.
• Formal-language interpretation should be made over this controlled grammar, not broad natural-language coverage.