Martin Schweinberger- Dr.
- Lecturer in Applied Linguistics at The University of Queensland
Martin Schweinberger
- Dr.
- Lecturer in Applied Linguistics at The University of Queensland
About
45
Publications
23,881
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
359
Citations
Introduction
I am Lecturer in Applied Linguistics at the University of Queensland. At UQ, I am Director of the Language Technology and Data Analysis Laboratory (LADAL) and I would consider myself a quantitative corpus linguist specialized in computational analyses of text and speech. In my research, I aim to combine and bridge the gap between computational linguistics and corpus linguistics.
Current institution
Additional affiliations
October 2018 - present
October 2018 - present
March 2018 - September 2018
Education
October 2008 - November 2011
August 2006 - February 2007
October 2000 - February 2008
Publications
Publications (45)
The focus of research on Singapore English has traditionally been on its structural features, while the relationship between English and other official languages of the Republic within the individual speaker has attracted much less interest, and comparatively little empirical data exist on the actual linguistic ecology of individual Singaporeans. T...
This paper uses a bibliometric analysis to map the field of Corpus Linguistics (CL) research in arts and humanities over the last 20 years, while tracking changes in the popular CL research topics, outlets, highly-cited authors, and geographical origins of published CL research. Based on a collection of the metadata of 5,829 CL-related articles fro...
This paper takes a corpus-based approach to study vulgar language in online communication across 20 English-speaking regions based on the Global Web-Based English Corpus (GloWbE). The identification of vulgarity combines word lists used in profanity detection with regular expressions to identify a wide range of vulgar elements including spelling va...
Topic models (TM), as introduced by Blei et al. (2003), are techniques used to infer linguistic themes from the co-occurrence of words in various textual entities, such as documents, paragraphs, sentences, or tweets. Traditionally, the process by which TMs derive these themes has been data-driven and unsupervised, meaning that human interpretative...
This volume illustrates new trends in corpus linguistics and shows how corpus approaches can be used to investigate new datasets and emerging areas in linguistics and related fields. It addresses innovative research questions, for example how prosodic analyses can increase the accuracy of syntactic segmentation, how tolerant English language teache...
In this paper we extend the discussion about reproducibility in corpus linguistics from quantitative to qualitative corpus-based approaches and argue that concerns about reproducibility can be addressed in interpretive research paradigms like corpus pragmatics. We first suggest that in interpretive research traditions, transparency is more importan...
Recent years have seen an increase in data and method reflection in corpus-based discourse analysis. In this article, we first take stock of some of the issues arising from such reflection (covering concepts such as triangulation, objectivity/subjectivity, replication, transparency, reflexivity, consistency). We then introduce a new ‘accountability...
This study considers a corpus‐based approach for analysing adjective amplifier systems ( very , really , so , etc.) in Hong Kong English (HKE), Indian Englsih (IndE) and Philippine English (PhiE) based on data from the International Corpus of English . The current study adds to existing research by providing insights into the understudied adjective...
This study analyzes the L1-acquisition of discourse like and its pragmatic functions in American English based on the Home-School Study of Language and Literacy Development component of the Child Language Data Exchange System (CHILDES). The data show that discourse like is already present in the speech of 3- and 4-year-old children and that even ve...
This study takes a corpus-based approach to investigating ongoing change in the Australian English adjective amplifier system based on the Australian component of the International Corpus of English (ICE). The paper analyzes changes in amplifiers across apparent time, with special attention being placed on amplifier–adjective–bigram frequencies, to...
While research on language change has formulated robust empirical generalisations about processes and motivations underlying the emergence and spread of linguistic elements, their decline and loss is less well understood. So far a systematic investigation into the processes and motivations of decline and loss in language change is lacking. This boo...
Information in turns-at-talk is hierarchically structured with some information made more prominent than other information. A key indicator of information structure is nuclear stress. In this paper we aim to identify factors that correlate with nuclear stress placement.
Based on exhaustive manual annotation of ten-word turns extracted from the demo...
Despite a marked shift towards blended learning in EAP due to massification of higher education and the impact of COVID-19, relatively little is known about student engagement with EAP genre education as realised through blended learning. This paper explores international undergraduate/graduate students’ perceptions of using an online text, image a...
Public discourse about the COVID-19 that appears on Twitter and other social media platforms provides useful insights into public concerns and responses to the pandemic. However, acknowledging that public discourse around COVID-19 is multi-faceted and evolves over time poses both analytical and ontological challenges. Studies that use text-mining a...
Data-driven learning (DDL), or the use of language corpora for the purposes of language learning and teaching, has seen a marked increase in research interest within ICT-rich WEIRD (Western, Educated, Industrialised, Rich and Democratic) contexts. However, less is known about its adoption in nations such as Indonesia where ICT/CALL training is unde...
This paper analyzes the use of very as an adjective amplifier by native speakers and advanced learners of English with diverse language backgrounds based on the International Corpus of Learner English (ICLE) and the Louvain Corpus of Native English Essays (LOCNESS). The study applies Multifactorial Prediction and Deviation Analysis Using Regression...
One of the possible explanations for the decline in IQ test results over the last few decades is the effect computerised digitisation has on our communicative behaviour. In what I call the digital discourse mode the content of what is said is not interpreted but dealt with as data, which are processed mechanically, resulting in a Yes/No, Right/Wron...
This study takes a corpus-based approach to examining co-occurrence patterns of amplifiers and adjectives based on the Irish and the New Zealand components of the International Corpus of English (ICE). The chapter investigates changes in amplifier-adjective-bigram frequencies, to provide insights into the mechanisms underlying lexical replacement....
This study aims to exemplify how language teaching can benefit from learner corpus research (LCR). To this end, this study determines how L1 and L2 English speakers with diverse L1 backgrounds differ with respect to adjective amplification, based on the International Corpus of Learner English (ICLE) and the Louvain Corpus of Native English Essays (...
Less is more? The impact of written corrective feedback on corpus-assisted L2 error resolution. Abstract The past decade has seen a sharp increase in research into L2 learners' direct use of language corpora (typically known as 'data-driven learning', DDL) for error resolution in L2 writing. However, a crucial yet underexplored variable in this pro...
This book showcases eleven studies dealing with corpora and the changing society. The theme of the volume reflects the fact that changes in society lead to changes in language and vice versa. Focusing on the English language, be it from Old English to the present, or a shorter time span in the immediate past, the contributors in this volume use a v...
This paper investigates the use of speech-unit final like (SUF like ) in standard Irish English (IrE) and takes a variationist approach based on the Irish component of the International Corpus of English (ICE-IRL). The analysis includes both sociolinguistic factors (age, gender, occupation type, religious affiliation, conversation type, audience si...
This study details a replicable method for annotating emotionality of natural language that can be used in sociopragmatic, corpus-based analyses of discourse. A case study uses a type of sentiment analysis based on the crowd-sourced Word-Emotion Association Lexicon to investigate the social stratification of emotives, i.e. words associated with one...
A corpus-based, quantitative analysis of ongoing change in the Australian English amplifier system.
Das Ziel des VowelChartProject (VCP) besteht in der Erstellung von personalisierten Vokaltrapezen für Studierende, die Englisch lernen. Außerdem soll das VCP auch den Nutzen der Sprachwissenschaft für Laien verdeutlichen. Das VCP wird bereits seit 2016 an der Universität Hamburg gefördert und soll nun auch für Studierende der Universität Kassel geö...
This corpus-based study examines the L1-acquisition of amplifiers in pre-adjectival slots in American English based on the Home-School Study of Language and Literacy Development (HSLLD) component of the Child Language Data Exchange System (CHILDES).
The aim of this study is to add to research which focuses on when and how variation is acquired by...
This paper investigates the use of eh in New Zealand English (NZE) and takes a quantitative, variationist approach based on the New Zealand component of the International Corpus of English (ICE New Zealand). The analysis includes both sociolinguistic (ethnicity, age, gender, occupation type) and psycholinguistic variables (priming). A mixed-effects...
Das Ziel des VowelChartProject (VCP) besteht in der Erstellung von personalisier- ten Vokaltrapezen (siehe Grafik rechts) für Studierende, die Englisch lernen. Das Projekt ist zur Zeit noch an der Universität Hamburg angesiedelt, aber soll bald auch für Studierende der Universität Kassel zugänglich gemacht werden. In einer früheren Projektphase wur...
This study takes a corpus-based approach to examining amplifying intensifiers across four genres in American English (AmE) using the Corpus of Historical American English (COHA) that comprises data from 1810 to 2000. From a language variation and change perspective intensifiers are particularly interesting as they play a crucial part in how speaker...
Among the core functions of language is to communicate one’s emotional state to others. Yet, there exists relatively little corpus-based research on how speakers convey their emotions to others. Furthermore, sociolinguistic research on emotional language which is based on corpus data is even more scare. To address this research gap, the current cor...
This study analyzes the effect of sociolinguistic variables on the frequency of swear words in Irish English based on the private dialogue section of the Irish component of the International Corpus of English (ICE). The results of mixed-effects regression models show that speakers between 19 and 33 are substantially more likely to use swear words c...
This study takes a corpus-based approach to examining co-occurrence patterns of amplifying intensifiers and adjectives (cf. 1) based on the New Zealand component of the International Corpus of English (ICE). From a language variation and change perspective intensifiers are particularly interesting as they play a crucial part in the “social and emot...
This study compares the use of like in Irish English (IrE) to its use in southeastern British English (SE-BrE). There are significant differences between the use of like in IrE and SE-BrE in terms of overall frequency, social meaning and positioning. This paper argues that the differences in the use of like require a functional explanation on two l...
The discourse marker LIKE is one of the most salient features of
present‐day English. Despite being deemed archaic, dismissed as meaningless and considered symptomatic of careless speech, this non‐standard feature has received
scholarly attention and attracted interest in the public media. In spite of being met with derision, its functional versati...
The present study examines the use of the discourse marker LIKE 1 in Northern Ireland. Previous claims concerning preferences for the employment of discourse marker LIKE based on the age and gender, as well as the acquisition of this marker are tested by employing a quantitative approach and statistical evaluation. The study takes a synchronic pers...
This study compares the use of like in Irish English (IrE) to its use in southeastern British English (SE-BrE). There are significant differences between the use of like in IrE and SE-BrE in terms of overall frequency, social meaning, and positioning. This paper argues that the differences in the use of like require a functional explanation on two...
Questions
Question (1)
Hi everyone,
I have used tree-based models (in my case Boruta) to streamline the model fitting process of mixed-effects modelling when dealing with many predictors.
In a comment to a recently submitted paper, a reviewer raised objections to this procedure based on Strobel, Malley & Tutz (2009) who state that "Note, however, that variable selection should not be conducted before applying another statistical method on the same learning data (Ambroise & McLachlan, 2002; Boulesteix et al., 2008; Leeb & Po¨tscher, 2006)."
However, immediately before they reject this procedure Strobel, Malley & Tutz (2009) write that "In addition to this, a black box method like random forests can be used to identify a small number of potentially relevant predictors from the full feature list, which can then be processed (e.g., by means of a familiar parametric method). This two-stage approach has been successfully applied in a variety of applications (see, e.g., Ward et al., 2006)."
I do not see this two-satge procedure as problematic but I am open to change my mind and wanted to ask you for your thoughts on this. So, what do you think?
Strobl, Carolin, James Malley, and Gerhard Tutz. 2009. "An introduction to recursive partitioning: rationale, application, and characteristics of classification and regression trees, bagging, and random forests." Psychological methods 14(4): 323–348.
