Building Analyses from Syntactic Inference in Local Languages: An HPSG Grammar Inference System

Abstract

We present a grammar inference system that leverages linguistic knowledge recorded in the form of annotations in interlinear glossed text (IGT) and in a meta-grammar engineering system (the LinGO Grammar Matrix customization system) to automatically produce machine-readable HPSG grammars. Building on prior work to handle the inference of lexical classes, stems, affixes and position classes, and preliminary work on inferring case systems and word order, we introduce an integrated grammar inference system that covers a wide range of fundamental linguistic phenomena. System development was guided by 27 geneologically and geographically diverse languages, and we test the system's cross-linguistic generalizability on an additional 5 held-out languages, using datasets provided by field linguists. Our system out-performs three baseline systems in increasing coverage while limiting ambiguity and producing richer semantic representations, while also producing richer representations than previous work in grammar inference.

20 years of the Grammar Matrix: cross-linguistic hypothesis testing of increasingly complex interactions

“…In particular, the MOM morphological inference system infers lexical entries and morphological rules for each form attested in the corpus of IGT data. These lexical entries and rules are merged into larger 51 Howell (2020) and Howell and Bender (2022) define the term grammar inference in the context of the AGGREGATION project; while earlier work such as Zamaraeva et al 2019a refers to grammar inference or describes it as a practice but does not provide a definition. [ 115 ] classes, resulting in grammars that account for a variety of previously unseen strings in a language.…”