Cifu: a Frequency Lexicon of Hong Kong Cantonese

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 3069–3077 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 3069 Cifu: a frequency lexicon of Hong Kong Cantonese Regine Lai, Grégoire Winterstein The Chinese University of Hong Kong, Université du Québec à Montréal Department of Linguistics and Modern Languages, Département de Linguistique ryklai@cuhk.edu.hk, winterstein.gregoire@uqam.ca Abstract This paper introduces Cifu, a lexical database for Hong Kong Cantonese (HKC) that offers phonological and orthographic information, frequency measures, and lexical neighborhood information for lexical items in HKC. The resource can be used for NLP applications and the design and analysis of psycholinguistic experiments on HKC. We elaborate on the characteristics and challenges specific to HKC that were relevant in the design of Cifu. This includes lexical, orthographic and phonological aspects of HKC, word segmentation issues, the place of HKC in written media, and the availability of data. We discuss the measure of Neighborhood Density (ND), highlighting how the analytic nature of Cantonese and its writing system affect that measure. We justify using six different variations of ND, based on the possibility of inserting or deleting phonemes when searching for neighbors and on the choice of data for retrieving frequencies. Statistics about the four genres (written, adult spoken, children spoken and child-directed) within the dataset are discussed. We find that the lexical diversity of the child-directed speech genre is particularly low, compared to a size-matched written corpus. The correlations of word frequencies of different genres are all high, but in general decrease as word length increases. Keywords: Hong Kong Cantonese, lexicon, lexical frequency, lexical neighborhood, neighborhood density 1. Introduction This paper describes the construction of Cifu: a lexical database for Hong Kong Cantonese.1 The resource lists lexical elements of Hong Kong Cantonese accompanied by a measure of their frequency in various genres, as well as morphological and orthographic information. The re- source is modeled after comparable ones for other lan- guages such as LEXIQUE3 for French (New and Pallier, 2019), or CELEX2 for English, Dutch and German (Baayen et al., 1995). In addition, the resource lists information about the “lexical neighborhood” of the items in the lexicon, i.e. other items in Cifu that are phonologically similar to the source item via some measuring distance (e.g. Levenshtein Distance). Such a resource is useful for both NLP applications (e.g. as the input of word segmenters, PoS taggers…) and for de- signing the materials of psycholinguistic experiments, typ- ically for balancing the frequency of lexical items, or in or- der to find candidate items for priming experiments. Ev- idence from psycholinguistics methodology such as prim- ing have shown that factors such as word frequency and phonological similarity to other words in the lexicon have a sizeable impact on language processing such as in word recognition and production, as well as in language acquisi- tion (e.g. (Dell and Gordon, 2003; Storkel et al., 2006; Vite- vitch and Luce, 1999)). These effects are found in popula- tions in a wide range …

Content truncated to 500 words