Negation of German polar words and expressions in automated analysis of text tonality

Abstract

The aim of the study is to describe various groups of negative words in the German language and the features of their functioning in text tonality analysis in the framework of automated natural language processing (NLP). Scientific analysis focuses on the ability of negative words to modify polar words and expressions, i.e., to change their connotation within an utterance. The study is novel in that it is the first to present in a systematic manner the syntactic-morphological relationship between certain groups of linguistic units of negation and polar words, which must be taken into account in NLP when carrying out sentiment analysis of German texts based on lexico-grammatical rules. As a result, all German language means of negation were divided into groups depending on the extent of their impact on the polar words in the sentence: tonality modification for only the adjoining polar element or the ability to interact with the parts of an utterance that are located at a distance of several tokens. For automatic tonality analysis of German texts, it is necessary to mark the morphological and syntactic features reflecting the peculiarities of both the selected groups of negation words and polar words and expressions at the stage of feature extraction.