The Semantic Comparison between Bare Noun Phrases in English and in Chinese

Since Carlson(1977), bare noun phrases(which willbe referred to as BNPs afterwards) have been a focus of linguistics in English. Gradually, the studies of bare noun phrase has also caught much attention in China. Despite of the extensive studies of BNPs in English and Chinese respectively, a systematic comparison between English and Chinese BNPs in semantic readings is still less developed. More detailed studies and comparisons are obviously desirable. Most linguists agree that the semantics of the predicates, syntactic distribution and pragmatic factors all have a role to play in determining the interpretation of BNPs. Thus the paper has systematically made a semantic comparison between English BNPs and Chinese BNPs in terms of different predicate types and different syntactic distributions following the four semantic readings of BNPs of Shen(2005), namely, generic reading, existential reading,definite reading and non-referential reading.

Hybrid Neural Network for Automatic Recovery of Elliptical Chinese Quantity Noun Phrases

In Mandarin Chinese,when the noun head appears in the context,a quantity noun phrase can be reduced to a quantity phrase with the noun head omitted.This phrase structure is called elliptical quantity noun phrase.The automatic recovery of elliptical quantity noun phrase is crucial in syntactic parsing,semantic representation and other downstream tasks.In this paper,we propose a hybrid neural network model to identify the semantic category for elliptical quantity noun phrases and realize the recovery of omitted semantics by supplementing concept categories.Firstly,we use BERT to generate character-level vectors.Secondly,Bi-LSTM is applied to capture the context information of each character and compress the input into the context memory history.Then CNN is utilized to capture the local semantics of n-grams with various granularities.Based on the Chinese Abstract Meaning Representation(CAMR)corpus and Xinhua News Agency corpus,we construct a hand-labeled elliptical quantity noun phrase dataset and carry out the semantic recovery of elliptical quantity noun phrase on this dataset.The experimental results show that our hybrid neural network model can effectively improve the performance of the semantic complement for the elliptical quantity noun phrases.