Computational Register Analysis and Synthesis

Argamon, Shlomo

doi:10.31235/osf.io/t64sy

Cited by 6 publications

(7 citation statements)

References 105 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The term 'register' defines the situational context of communication (Biber & Conrad 2005). Registers are thus primarily differentiated in non-linguistic terms, although linguistic variability can be observed when comparing different registers with each other (Argamon 2019;Biber & Conrad 2005). Contributing factors to register variation (amongst others) are situational characteristics such as the setting and channel of communication, discourse participants and their relationships, as well as the purpose of communication (Biber & Conrad 2019), which might require certain linguistic features in order to be successful.…”

Section: Defining Registermentioning

confidence: 99%

The medium is not the message

Scheffler

Kern

Seemann

2022

View full text Add to dashboard Cite

Linguistic expressions in social media vary along many axes, including author style, the specific medium and its affordances, and others. In this paper, we argue that different registers must be distinguished within social media and that register should be included as an important factor independent of (social) medium in analyses of variable linguistic phenomena. We introduce a new German cross-media corpus, consisting of blog posts and tweets from the same 44 authors. We define the registers as ‘Informative’, ‘Narrative’, and ‘Persuasive’, based on situational characteristics of the texts. We then correlate the registers with two variable linguistic phenomena: German modal and intensifying particles. In each case, we document considerable inter- and intraindividual variation in the expressions used and their frequency across texts. The statistical analysis shows that the register grouping corresponds more closely to linguistic similarities between texts than the grouping by medium does.

show abstract

Section: Defining Registermentioning

confidence: 99%

The medium is not the message

Scheffler

Kern

Seemann

2022

View full text Add to dashboard Cite

show abstract

“…This section describes the key differences of this study from other studies in Automatic Genre Identification (AGI) and Multi-Dimensional Analysis (MDA). The main aim of AGI studies is to predict stylistic properties using extractable features (Santini et al, 2010;Argamon, 2019). Experiments have been carried out with such features as POS tags (Karlgren and Cutting, 1994), broader linguistic features (Kessler et al, 1997), including theoretically motivated features from Systemic-Functional Linguistics (Argamon et al, 2007).…”

Section: Related Studies On Computational Analysis Of Genresmentioning

confidence: 99%

Genre annotation for the Web

Sharoff

2021

View full text Add to dashboard Cite

This paper describes a digital curation study aimed at comparing the composition of large Web corpora, such as enTenTen, ukWac or ruWac, by means of automatic text classification. First, the paper presents a Deep Learning model suitable for classifying texts from large Web corpora using a small number of communicative functions, such as Argumentation or Reporting. Second, it describes the results of applying the automatic classification model to these corpora and compares their composition. Finally, the paper introduces a framework for interpreting the results of automatic genre classification using linguistic features. The framework can help in comparing general reference corpora obtained from the Web and in comparing corpora across languages.

show abstract

“…It has been shown that the use of topic modelling helps in finding the differences between the Web corpora (Fothergill et al, 2016;Sharoff, 2013). Since the arrival of machine learning methods in the 1990s, genre classification and related approaches to classification of texts with respect to their stylistic features developed from (Karlgren and Cutting, 1994) to (Pritsos and Stamatatos, 2018), see a recent overview in (Argamon, 2019). However, genre classification methods have not been yet applied to very large corpora from the Web.…”

Section: Related Studiesmentioning

confidence: 99%

Know thy corpus! Robust methods for digital curation of Web corpora

Sharoff¹

2020

Preprint

View full text Add to dashboard Cite

This paper proposes a novel framework for digital curation of Web corpora in order to provide robust estimation of their parameters, such as their composition and the lexicon. In recent years language models pre-trained on large corpora emerged as clear winners in numerous NLP tasks, but no proper analysis of the corpora which led to their success has been conducted. The paper presents a procedure for robust frequency estimation, which helps in establishing the core lexicon for a given corpus, as well as a procedure for estimating the corpus composition via unsupervised topic models and via supervised genre classification of Web pages. The results of the digital curation study applied to several Web-derived corpora demonstrate their considerable differences. First, this concerns different frequency bursts which impact the core lexicon obtained from each corpus. Second, this concerns the kinds of texts they contain. For example, OpenWebText contains considerably more topical news and political argumentation in comparison to ukWac or Wikipedia. The tools and the results of analysis have been released.

show abstract

Computational Register Analysis and Synthesis

Cited by 6 publications

References 105 publications

The medium is not the message

The medium is not the message

Genre annotation for the Web

Know thy corpus! Robust methods for digital curation of Web corpora

Contact Info

Product

Resources

About