Learning the Ordering of Coordinate Compounds and Elaborate Expressions in Hmong, Lahu, and Chinese

21138

Authors

nan

Abstract

AbstractCoordinate compounds (CCs) and elaborate expressions (EEs) are coordinate constructions common in languages of East and Southeast Asia. Mortensen (2006) claims that (1) the linear ordering of EEs and CCs in Hmong, Lahu, and Chinese can be predicted via phonological hierarchies and (2) that these phonological hierarchies lack a clear phonetic rationale. These claims are significant because morphosyntax has often been seen as in a feed-forward relationship with phonology, and phonological generalizations have often been assumed to be phonetically “natural”. We investigate whether the ordering of CCs and EEs can be learned empirically and whether computational models (classifiers and sequence-labeling models) learn unnatural hierarchies similar to those posited by Mortensen (2006). We find that decision trees and SVMs learn to predict the order of CCs/EEs on the basis of phonology, beating strong baselines for all three languages, with DTs learning hierarchies strikingly similar to those proposed by Mortensen. However, we also find that a neural sequence labeling model is able to learn the ordering of elaborate expressions in Hmong very effectively without using any phonological information. We argue that EE ordering can be learned through two independent routes: phonology and lexical distribution, presenting a more nuanced picture than previous work.

Keywords

nan

Full Content

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 3656 - 3669 July 10-15, 2022 ©2022 Association for Computational Linguistics Learning the Ordering of Coordinate Compounds and Elaborate Expressions in Hmong, Lahu, and Chinese Chenxuan Cui Katherine J. Zhang Language Technologies Institute, Carnegie Mellon University cxcui@cs.cmu.edu, kjzhang@cmu.edu, dmortens@cs.cmu.edu David R. Mortensen Abstract Coordinate compounds (CCs) and elaborate expressions (EEs) are coordinate constructions common in languages of East and Southeast Asia. Mortensen (2006) claims that (1) the lin- ear ordering of EEs and CCs in Hmong, Lahu, and Chinese can be predicted via phonological hierarchies and (2) these phonological hierar- chies lack a clear phonetic rationale. These claims are significant because morphosyntax has often been seen as in a feed-forward re- lationship with phonology, and phonological generalizations have often been assumed to be phonetically “natural”. We investigate whether the ordering of CCs and EEs can be learned empirically and whether computa- tional models (classifiers and sequence label- ing models) learn unnatural hierarchies similar to those posited by Mortensen (2006). We find that decision trees and SVMs learn to predict the order of CCs/EEs on the basis of phonol- ogy, with DTs learning hierarchies strikingly similar to those proposed by Mortensen. How- ever, we also find that a neural sequence la- beling model is able to learn the ordering of elaborate expressions in Hmong very effec- tively without using any phonological informa- tion. We argue that EE ordering can be learned through two independent routes: phonology and lexical distribution, presenting a more nu- anced picture than previous work. [ISO 639-3: hmn, lhu, cmn] 1 1 Introduction In many languages of East and Southeast Asia, there are common constructions in which two words or phrases are coordinated without an overt marker like a conjunction (Hanna, 2013; Filbeck, 1996; Johns and Strecker, 1987; Wheatley, 1982; Matisoff, 1973; Pan and Cao, 1972; Watson, 1966; Banker, 1964). In coordinate compounds (CCs), two words are combined to form a compound word 1Code and data available at: https://github.com/ dmort27/elab-order whose semantics are often a generalization of those of the two conjoined words. Elaborate expressions (EEs) are similar, except that they can consist of two phrases (rather than words) and include a re- peated word. Take the following examples: (1) Chinese coordinate compounds (CCs) 父母 fùmǔ father-mother ‘parents’ 花木 huāmù flower-tree ‘vegetation’ 天地 tiāndì heaven-earth ‘universe’ 国家 guójiā country-home ‘nation’ 风水 fēngshuǐ wind-water ‘geomancy’ (2) Lahu elaborate expressions (EEs) a. ɔ̂ four cē corner ɔ̂ four phɔ̂ side ‘at every corner’ b. chɔ people phôʔ pile chɔ people dì lump ‘a throng of people’ c. câ eat cûʔ scarce dɔ̀ drink cûʔ scarce ‘have nothing to eat or drink’ Coordinating compounds are found throughout the world, with varying semantic relationships be- tween the whole and the parts (Obermüller, 2015). Elaborate expressions are most common in main- land Southeast Asia, where they occupy a position of great prominence. They are often associated with elevated styles of discourse, but they …

Content truncated to 500 words

Labeled Text

AbstractCoordinate compounds (CCs) and elaborate expressions (EEs) are coordinate constructions common in languages of East and Southeast Asia. Mortensen (2006) claims that (1) the linear ordering of EEs and CCs in Hmong, Lahu, and Chinese can be predicted via phonological hierarchies and (2) that these phonological hierarchies lack a clear phonetic rationale. These claims are significant because morphosyntax has often been seen as in a feed-forward relationship with phonology, and phonological generalizations have often been assumed to be phonetically “natural”. We investigate whether the ordering of CCs and EEs can be learned empirically and whether computational models (classifiers and sequence-labeling models) learn unnatural hierarchies similar to those posited by Mortensen (2006). We find that decision trees and SVMs learn to predict the order of CCs/EEs on the basis of phonology, beating strong baselines for all three languages, with DTs learning hierarchies strikingly similar to those proposed by Mortensen. However, we also find that a neural sequence labeling model is able to learn the ordering of elaborate expressions in Hmong very effectively without using any phonological information. We argue that EE ordering can be learned through two independent routes: phonology and lexical distribution, presenting a more nuanced picture than previous work.

Publication Details

Paper ID:
21138
Year:
2022
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan