Computational Modeling of Affixoid Behavior in Chinese Morphology

Proceedings of the 28th International Conference on Computational Linguistics, pages 2879–2888 Barcelona, Spain (Online), December 8-13, 2020 2879 Computational Modeling of Affixoid Behavior in Chinese Morphology Yu-Hsiang, Tseng Graduate Institute of Linguistics National Taiwan University seantyh@gmail.com Shu-Kai Hsieh Graduate Institute of Linguistics National Taiwan University shukaihsieh@ntu.edu.tw Pei-Yi, Chen Graduate Institute of Linguistics National Taiwan University peiyijessychen@gmail.com Sara Court Graduate Institute of Linguistics National Taiwan University sarakcourt@gmail.com Abstract The morphological status of affixes in Chinese has long been a matter of debate. How one might apply the conventional criteria of free/bound and content/function features to distinguish word-forming affixes from bound roots in Chinese is still far from clear. Issues involving polysemy and diachronic dynamics further blur the boundaries. In this paper, we propose three quantitative features in a computational model of affixoid behavior in Mandarin Chinese. The results show that, except for in a very few cases, there are no clear criteria that can be used to identify an affix’s status in an isolating language like Chinese. A diachronic check using contextualized embeddings with the WordNet Sense Inventory also demonstrates the possible role of the polysemy of lexical roots across diachronic settings. 1 Introduction In typological terms, Mandarin Chinese is a canonical isolating language. Unlike other more agglutinative languages with rich derivational morphology and large numbers of highly produc- tive affixes, Mandarin words most often consist of either monosyllabic roots or polysyllabic, multi-root compounds (Arcodia, 2012; Huang et al., 2018). Traditionally, a derivational affix is defined as a bound morpheme that attaches to a root or stem to change its syntactic category or contribute additional semantic content (Lieber, 2017). However, the status of derivational affixes in Chinese and the morphological processes that characterize them is still largely a matter of debate. Morphological derivation and compounding are often contrasted with one another. Despite this, defining a clear cross-linguistic distinction between the two processes can be troublesome. In languages such as English, a distinction between free and bound morphemes provides a relatively straightforward benchmark. A suffix like “-ology” is bound in English and must always attach to a root in order to productively derive new words in the language. On the other hand, free morphemes in English may combine together to form a compound, such as “green” and “house” in the English word “greenhouse”. In Mandarin, however, issues arise when free/bound and content/function criteria are used to distinguish word-forming affixes from bound roots, as these features are far from clearly defined in the language (Huang et al., 2018; Packard, 2000). Issues involving polysemy and diachronic change further blur the boundaries (Arcodia, 2014; Bauer, 2005; Huang et al., 2018; Ralli, 2010). This work is licensed under a Creative Commons Attribution 4.0 International License. License details: http: //creativecommons.org/licenses/by/4.0/. 2880 The morphological status of most Chinese morphemes is indeterminate, which leads some to conclude that there is an extreme poverty of affixation in Chinese morphology; it is rarely derivational and hardly ever inflectional. Previous studies suggest so-called “pure” or “bona fide” affixes display high morphological productivity …

Content truncated to 500 words