Collocations play a very important role in language description, especially in identifying meanings of words. Modern lexicography’s inevitable part of meaning deduction are lists of collocates ranked ...by some statistical measurement. In the paper, we present a comparison between two approaches to the ranking of collocates: (a) the logDice method, which is dominantly used and frequency-based, and (b) the fastText word embeddings method, which is new and semantic-based. The comparison was made on two Slovene datasets, one representing general language headwords and their collocates, and the other representing headwords and their collocates extracted from a language for special purposes corpus. In the experiment, two methods were used: for the quantitative part of the evaluation, we used supervised machine learning with the area-under-the-curve (AUC) ROC score and support-vector machines (SVMs) algorithm, and in the qualitative part the ranking results of the two methods were evaluated by lexicographers. The results were somewhat inconsistent; while the quantitative evaluation confirmed that the machine-learning-based approach produced better collocate ranking results than the frequency-based one, lexicographers in most cases considered the listings of collocates of both methods very similar.
This article studies the phrasemes comprising an ethnonym in the source language (French) as well as the target language (Slovak). This approach is contrastive and the phrasemes have been classified ...according to the type of equivalence (total equivalent, partial equivalent and phrasemes without equivalent). The aim of the research was to analyse 27 phrasemes with the help of the corpus linguistics method (relative frequency and logDice association measure), and four monolingual corpora (the corpora frTenTen12, skTenTen11, Emolex, prim-7.0-public-all) with approx. 130 million up to approx. 10 billion words in each of them, so it is a fairly wide range of language materials. Firstly, we focus on the current state of French and Slovak phraseology. We present the distribution of phrasemes into three types: general, professional and so-called mixed (of which the last type represents our own proposition). Then, by translating the source language-culture into the target language-culture, we demonstrate the three basic types of phrasemes equivalence but our attention is on the first two types. Afterwards, we present quantitative methods of corpus linguistics (four monolingual corpora, relative frequency and the logDice association measure). Then, we analyse 27 specific phrasemes. This is qualitative analysis (their distribution into three types of equivalence as well as their repair in general, professional and mixed phrasemes), but also quantitative analysis (analysis based on relative frequency and also on logDice association measure). In the end, we demonstrate and evaluate the results of our research. The research objectives are set to find out the frequency of phrasemes in various types of texts and the level of their specificity within the framework of each of the corpora, based on which it is possible to propose which of the phrasemes should be placed at the forefront for looking up an entry and its individual components or a phraseme as a whole, and thus contribute to supporting the creation of current French/Slovak lexicography and phraseography. Moreover, from the point of view of teaching foreign languages, we can use the second type of equivalence as a contrasting factor between French and Slovak language-cultures because they can either easily interfere with other phrasemes of the target language-culture or be not well understood in the target language-culture.
PDF
