Modeling Communicative Purpose with Functional Style: Corpus and Features for German Genre and Register Analysis

Abstract

While there is wide acknowledgement in NLP of the utility of document characterization by genre, it is quite difficult to determine a definitive set of features or even a comprehensive list of genres. This paper addresses both issues. First, with prototype semantics, we develop a hierarchical taxonomy of discourse functions. We implement the taxonomy by developing a new text genre corpus of contemporary German to perform a text based comparative register analysis. Second, we extract a host of style features, both deep and shallow, aiming beyond linguistically motivated features at situational correlates in texts. The feature sets are used for supervised text genre classification, on which our models achieve high accuracy. The combination of the corpus typology and feature sets allows us to characterize types of communicative purpose in a comparative setup, by qualitative interpretation of style feature loadings of a regularized discriminant analysis. Finally, to determine the dependence of genre on topics (which are arguably the distinguishing factor of sub-genre), we compare and combine our style models with Latent Dirichlet Allocation features across different corpus settings with unstable topics.

Exploring the role of lexis and grammar for the stable identification of register in an unrestricted corpus of web documents

“…These included both topical features associated with, e.g., traveling, and specific lexical items that could not be captured with grammatical tags. Although a detailed analysis of all the registers is out of the possibilities of the current study, this finding does shed light also on the relationship between topics and registers discussed in many register identification studies (e.g., Asheghi et al 2014; see also Haider and Palmer 2017). Specifically, our analysis suggested that topics are relevant for some registers, such as Travel blog, for which lexical information has high predictive importance, while for others, such as Interview, this is not the case.…”

Beyond English: Considering Language and Culture in Psychological Text Analysis

“…Another issue concerns the specificity of text samples on which validity and equivalence tests were performed. In this sense, the communication context (text type, genre, register) is an important factor that can produce substantial variation both in the frequency of language features and in the associations with other variables, especially psychological ones (Pennebaker et al, 2007;Daems et al, 2013;Haider and Palmer, 2017;Biber and Conrad, 2019;Kučera et al, 2020;Dudãu and Sava, 2021). Differences in the sensitivity of LIWC for detecting psychological markers in different types of text (English only), were shown in the meta-analysis of Chen et al (2020), in which, for example, the strength of the relationship between extraversion and positive emotion words varied significantly and substantially across communication contexts (e.g., asynchronous/synchronous and public/private communication).…”